The influence of thermal history on
hydrocarbon prospectivity in the Central Irish Sea Basin

 Paul F. Green¹, Ian R. Duddy¹, Richard J. Bray²,
William I. Duncan³ and Dermot Corcoran⁴

¹Geotrack International, 37 Melville Road, West Brunswick, Victoria 3055, Australia
²Geotrack UK Office, 5 The Linen Yard, South St, Crewkerne, Somerset TA18 7HJ, UK
³ Veba Oil and Gas UK Ltd, Bowater House, 114 Knightsbridge, London SW1X 7LD, UK
⁴ Department of Geology, Trinity College, Dublin 2, Ireland

Abstract

Thermal history reconstruction studies of four hydrocarbon exploration wells located in the Central Irish Sea Basin (CISB) reveal three major episodes of heating and cooling, each of regional extent. Units throughout the pre-Quaternary section intersected in wells 42/12-1, 42/16-1 and 42/17-1 began to cool from their maximum post-depositional palaeotemperatures in the Early Cretaceous, between 120 and 115 Ma. Cooling from subsequent palaeotemperature peaks began in the Late Cretaceous-Early Tertiary (between 70 and 55 Ma) and Late Tertiary (25-0 Ma). Results from well 42/21-1 are dominated by the two more recent episodes, and show no evidence of the Early Cretaceous episode. This is thought to reflect a different structural setting of this well, within a North Celtic Sea – Cardigan Bay trend.

Palaeotemperature profiles suggest that heating in each episode was due largely to deeper burial, with subsequent cooling due mainly to uplift and erosion. A maximum of around 3 km of additional section (of Late Triassic to Early Cretaceous age) is required to explain the observed Early Cretaceous palaeotemperatures. Appropriate values for the Late Cretaceous-Early Tertiary and Late Tertiary episodes are ~2 km and ~1 km, respectively.

All of these cooling episodes correlate closely with similar episodes recognised from previous studies in surrounding regions, from onshore Ireland, Scotland, South Wales and Northern, Eastern, Central and SW England, and each appears to be of truly regional extent.

Exploration risk in the CISB generation can be significantly reduced through recognition of the major palaeo-thermal episodes which have affected the region, and the variation in the magnitude of their effects across the region. The challenge for future exploration in the region is to identify regions where the main phase of hydrocarbon generation post-dated structuring.

Introduction

Understanding the timing of hydrocarbon generation is a critical aspect of assessing regional hydrocarbon prospectivity. This is particularly important in sedimentary basins which have undergone a series of palaeo-thermal episodes, as a result of which a given source rock horizon may have reached maximum maturity at different times in various locations across the basin. The time at which a particular source rock cools from its maximum palaeotemperature effectively defines the time at which active hydrocarbon generation ceases. The relationship between this and the time at which various traps were formed can exert a critical control on hydrocarbon prospectivity, since only those traps formed prior to the main phase of generation will be available to be charged at that time. Accurate reconstruction of the thermal history of source rock sequences is therefore of major importance in reducing exploration risk in such regions. In addition, recognition of later tectonic episodes which might lead to breaching of seals, remigration and loss of charge, is another important aspect of hydrocarbon prospectivity which can be investigated through thermal history studies.

In marked contrast to the East Irish Sea Basin which contains significant hydrocarbon reserves (Colter, 1997), the history of hydrocarbon exploration in the Central Irish Sea Basin has been disappointing, despite many similarities in the geology of the two basins. While a number of factors may be responsible for the differences in hydrocarbon prospectivity between these two provinces, here we focus on the role of thermal history of potential source rocks.

 

In the Irish Sea region in general, source rocks are recognised within the Namurian section (Hardman et al., 1993; Armstrong et al. 1995, 1997), while Jurassic source rocks are also developed in surrounding regions (e.g. Scotchman and Thomas, 1995). The area is characterised by a series of potential structuring episodes, of end-Carboniferous, Mesozoic and Tertiary age (e.g. Stuart and Cowan, 1991; Maddox et al, 1995). Previous thermal history studies from areas adjacent to the CISB, including the East Irish Sea Basin (Green et al., 1997) and onshore Ireland (Green et al. 1998) have revealed a series of palaeo-thermal episodes, in the Late Carboniferous, Jurassic, Early Cretaceous, Early Tertiary and Late Tertiary, with the magnitude of peak palaeotemperatures during each episode varying significantly across these regions. The similarity in results from these regions, located to the east and west of the Central Irish Sea, suggests that the interplay between these various palaeo-thermal episodes and various structuring events is also likely to be crucial in understanding the history of hydrocarbon generation and accumulation in the CISB.

The Central Irish Sea Basin

The Central Irish Sea Basin (CISB) consists of a NE-SW trending, Late Palaeozoic-Cenozoic, transtensional half-graben system which has experienced a multiphase inversion history. The basin is bounded to the north by the Mid Irish Sea Uplift and to the south it is separated from the St. George's Channel-Cardigan Bay Basin by the offshore extension of St. Tudwal's Arch (Figure 1).

Five exploration wells have been drilled to date in the Irish sector of the CISB, all of which have been plugged and abandoned as dry holes. Potential hydrocarbon source rocks of Westphalian and Liassic ages have been recognised in the Central Irish Sea area (Corcoran and Clayton, 1999) although the efficacy of the Westphalian source rock system in the Central Irish Sea Basin has been questioned (Floodpage et al. 1999). The primary exploration target has been the Lower Triassic Sherwood Sandstone Group, sealed by evaporites and shales of the Upper Triassic Mercia Mudstone Group and sourced by Carboniferous shales and coals. Three of these wells appear to have tested valid hydrocarbon traps (Floodpage et al., 1999), strengthening the suggestion that the timing of hydrocarbon charge is a major exploration risk factor in the CISB.

Figure 1: Location map of the Central Irish Sea Basin and adjacent regions, showing locations of hydrocarbon exploration wells from which samples were analysed for this study. Stratigraphic columns for each well show the breakdown of the section intersected in each well.

However, maturation modelling of potential source rock horizons in the CISB is hampered by the severely truncated rock record. The challenge of thermal history reconstruction is to offer constraints to the thermal evolution of these potential source rock horizons in a multiphase inversion setting.

With this in mind, we report here results from four CISB wells (Figure 1), forming part of an ongoing study designed to determine the thermal history of potential hydrocarbon source rocks across the region. Results from two of these wells have been published previously (Duncan et al., 1998), but some aspects of these results have been reassessed, while comparison with new data provides tighter constraints on the interpretation of these older data than was previously possible.

Thermal history reconstruction using AFTA^® and VR

Thermal History Reconstruction (THR) is based on application of Apatite Fission Track Analysis (AFTA) and vitrinite reflectance (VR). Using THR, we can identify the timing of dominant episodes of heating and cooling that have affected a sedimentary section, quantify the palaeotemperatures through the section, and characterise mechanisms of heating and cooling (as described in detail by e.g. Bray et al., 1992; Duddy et al,. 1994; Green et al., 1995).

AFTA is based on analysis of radiation damage trails ("fission tracks") within the crystal lattice of detrital apatite grains, which are a common constituent of most sandstones and coarser sediments. The continuous production of new fission tracks through time, coupled to the reduction in track length as a function of temperature and time, provides the basis of the technique (Green et al., 1989a, 1989b). As temperature increases, track length is progressively reduced, due to partial repair of the radiation damage constituting the tracks ("partial annealing"). Once the temperature reaches some maximum value and begins to decrease, track length is essentially "frozen" at the value reached at the thermal maximum. Temperature dominates over time in the kinetics of this process, such that an increase of 10°C produces a similar change in length as an order of magnitude increase in time. Thus, most tracks are reduced to the same length regardless of when they formed. A sample which reached a maximum palaeotemperature of say 90 or 100°C at some time in the past, and then cooled and resided at lower temperatures until the present day, will therefore contain two populations of tracks: a shorter component which represents tracks formed up to the time at which cooling began, and a longer component representing tracks formed after the onset of cooling. The length of the shorter component is diagnostic of the maximum palaeotemperature reached prior to the onset of cooling, while that of the longer component is controlled by the history after the onset of cooling. The proportion of short to long tracks provides information on the time at which cooling began. For example, early cooling would result in few short tracks and mostly long tracks, while more recent cooling will produce more shorter lengths and fewer longer tracks. Thus, analysis of the distribution of track lengths provides estimates of both the time of cooling and the magnitude of the maximum palaeotemperature.

The number of tracks in a polished surface can also be used to measure a "fission track age". In the absence of significant length reduction, this parameter would measure the time over which tracks have been retained. But because the probability of a track intersecting a surface depends on the track length, when length reduction is sufficiently severe, the fission track age is significantly reduced, and must be interpreted together with track length data in terms of thermal history rather than as an indicator of the timing of a discrete event.In samples which reached sufficiently high values of maximum palaeotemperature, the track length is reduced to zero, because all of the radiation damage constituting the track is totally repaired ("total annealing"). Such samples only begin to retain tracks after cooling below this critical limit, which is typically around 110 or 120°C, depending on heating rate and apatite composition (chlorine content). In such cases only a minimum limit is possible on the magnitude of the maximum palaeotemperature, but the fission track age (combined with track length data which record the post-cooling thermal history) provides key information on the time of cooling.Thermal history interpretation of AFTA and VR data is based on a detailed knowledge of the kinetic responses of both systems, which are well calibrated from studies in both geological and laboratory conditions.

Thermal history information is extracted from the AFTA data by modelling measured AFTA parameters (fission track age and track length distributions) through a variety of possible thermal history scenarios, varying the magnitude and timing of the maximum palaeotemperature in order to define the range of values of each parameter which give predictions consistent with the measured data within 95% confidence limits. The basics of this modelling procedure are well established for mono-compositional apatites (e.g. Green et al. 1989b), based on a series of laboratory experiments on Durango apatite (Green et al. 1986; Laslett et al. 1987; Duddy et al. 1988). However, the annealing kinetics of fission tracks in apatite are known to be affected by the chlorine content (Green et al. 1986), and in the studies described here, thermal history solutions have been extracted from the AFTA data using a "multi-compositional" kinetic model which makes full quantitative allowance for the effect of Cl content on annealing rates of fission tracks in apatite (Green et al. 1996). This model is calibrated using a combination of laboratory and geological data from a variety of sedimentary basins around the world. Palaeotemperature estimates from AFTA are quoted as a range (corresponding to ±95% confidence limits) and have an absolute uncertainty of between ±5 and ±10°C.

Observed VR values are converted to maximum palaeotemperatures using the kinetic model developed by Burnham and Sweeney (1989) and Sweeney and Burnham (1990). Information on the timing of these maximum palaeotemperatures is provided by the AFTA data. The VR derived palaeotemperature estimates are shown as single values but probably have a precision between 5 and 10°C. The kinetic response of vitrinite reflectance as described by Burnham and Sweeney (1989) is very similar to the fission track annealing kinetic model developed by Laslett et al. (1987) to describe the kinetics of fission track annealing in Durango apatite. Total fission track annealing in apatites with typical Cl content corresponds to a VR value of ~0.7%, regardless of heating rate (Duddy et al. 1991, 1994).

Unlike VR data, AFTA data also provide some control on the history after cooling from maximum palaeotemperatures, through the lengths of tracks formed during this period. Wherever possible, AFTA data from each sample are interpreted in terms of two episodes of heating and cooling, using assumed heating and cooling rates during each episode, with the maximum palaeotemperature reached during the earlier episode. The timing of the onset of cooling and the peak palaeotemperatures during the two episodes are varied systematically, and by comparing predicted and measured parameters the range of conditions which are compatible with the data can be defined. One additional episode during the cooling history is normally the limit of resolution from typical AFTA data. (In rare instances, such as in one sample from the 42/16-1 well described in this paper, information on three discrete episodes may be obtained from a single sample.)

Palaeotemperature profiles, palaeogeothermal gradients and removed section

Analysis of a series of samples using AFTA and VR over a range of depths reveals the variation of maximum palaeotemperature with depth - the "palaeotemperature profile" characterising each episode. From data which reveal multiple episodes of heating and cooling, separate palaeotemperature profiles can often be constructed for each episode (see e.g. the discussion of Inner Moray Firth well 12/16-1 in Green et al. 1995).

The form of the palaeotemperature profile characterising a particular palaeo-thermal episode provides key information on likely mechanisms of heating and cooling in that episode. Heating due solely to deeper burial should produce a more or less linear palaeotemperature profile with a similar gradient to the present temperature profile. In contrast, heating due primarily to increased basal heat flow (perhaps also with a minor component of deeper burial) should produce a more or less linear palaeotemperature profile with a higher gradient than the present temperature profile.

Heating due to the passage of hot fluids can produce a variety of non-linear palaeotemperature profiles, with different forms depending on the timescale of heating (see e.g. Ziagos and Blackwell, 1986; Duddy et al., 1994). Heating effects due to minor igneous intrusions can produce purely local anomalies, or may be more widespread if they cause circulation of heated fluids on a regional scale (e.g. Summer and Verosub 1989).

In sections where heating was due to deeper burial, either alone or possibly combined with elevated heat flow, fitting a line to the palaeotemperature profile provides an estimate of the palaeogeothermal gradient. Extrapolating this to an assumed palaeo-surface temperature then provides an estimate of the amount of section removed by erosion (this analysis depends critically on several assumptions, as discussed by Bray et al. 1992). These two parameters are highly correlated, such that higher palaeogeothermal gradients require correspondingly lower values of removed section, and vice versa. Statistical techniques allow definition of the range of each parameter allowed by the palaeotemperature constraints within 95% confidence limits (Bray et al., 1992).

In summary, AFTA data are used to identify the timing of major cooling episodes, while AFTA and VR data provide estimates of the magnitude of maximum or peak palaeotemperatures in each episode. From these results, palaeotemperature-depth profiles are constructed for each episode, and these provide insight into the mechanism of heating and cooling. If these profiles are linear, palaeogeothermal gradients can be determined, and where appropriate, former depths of burial can be estimated. Non-linear profiles must reflect processes not directly related to depth of burial.

Results from Central Irish Sea Basin wells

Identification of palaeo-thermal episodes

Figure 2 illustrates the range of timing for the onset of cooling derived from AFTA data in samples from three of the four CISB wells analysed in this study. In each well, this timing information is compared with the variation in stratigraphic age through the well. Two samples analysed from well 42/17-1 failed to yield any apatite. In most of the samples, the AFTA data require at least two episodes of cooling, as shown, while as mentioned previously, the shallower sample analysed from well 42/16-1 shows very clear evidence of three distinct cooling episodes.

Figure 2: Timing constraints derived from AFTA in individual samples from three CISB wells. Vertical shaded bars highlight the range of timings with which data from all samples in each well are consistent. Data from all three wells define three synchronous events, in which cooling began between 120 and 115 Ma, 70 and 55 Ma and 25 and 0 Ma. Note that AFTA data from the shallower sample in well 42/16-1 define three separate cooling episodes, while other samples define two or one episodes. This depends on the quality of the AFTA data, the magnitude of peak palaeotemperatures in individual episodes and the spread of chlorine contents in apatites from each sample.

The vertical bars in each plot highlight the range of timing consistent with all samples from each well. Given the relatively close proximity of these three wells, it seems reasonable to assume that the palaeo-thermal effects recognised in each well represent synchronous events. On this basis, inspection of Figure 2 shows that results from all three wells can be explained in terms of three palaeo-thermal episodes, with cooling beginning in the Early Cretaceous (some time between 120 and 115 Ma), Late Cretaceous - Early Tertiary (between 70 and 55 Ma) and Late Tertiary (between 25 and 0 Ma). We emphasise that the quoted ranges refer to the time at which cooling began, and it is not implied either that all cooling in each episode occurred within each interval or that cooling necessarily encompassed the entire interval.

Quantification of palaeotemperatures in individual episodes

Figure 3 shows palaeotemperatures derived from AFTA and VR in individual samples from the four wells, plotted against sample depth (with respect to KB in each well). Also shown are present-day thermal gradients derived from corrected BHT data in each well. Results from the four wells show very similar features, as highlighted in the following discussion of each well.

Well 42/12-1: Results from this well were originally described Duncan et al. (1998) and the results shown in Figure 3 are as reported there. In most of the AFTA samples, all tracks were totally annealed prior to the Early Cretaceous cooling episode, and provide only minimum estimates of the maximum palaeotemperature in this episode. AFTA data also provide estimates of Late Cretaceous to Early Tertiary palaeotemperatures, as shown. VR data from the Carboniferous section in this well are between 1.77% and 1.95%, suggesting maximum palaeotemperatures in the range 160 to 200°C. Since AFTA data provide only minimum estimates, which are ~40 to 80°C less than the maximum values derived from VR, it is not immediately clear whether the maximum palaeotemperatures indicated by the VR data were attained during the Early Cretaceous episode revealed by AFTA or possibly during an earlier episode. Circumstantial evidence supporting an Early Cretaceous maximum comes from the observation that a linear profile with a similar gradient to the present-day temperature profile can satisfy all the AFTA-based palaeotemperature constraints and also those from VR as illustrated in Figure 3. Thus, if the VR data were to represent maximum temperatures reached in an earlier (pre-Cretaceous) episode, the Early Cretaceous palaeo-thermal episode would have to be described by a much lower palaeogeothermal gradient, for which no support is found in any of the results from other wells. As will be discussed, the consistency of results from this well with those from the 42/16-1 and 42/17-1 wells strongly support the conclusion that the VR data in this well represent maximum palaeotemperatures reached during the Early Cretaceous episode, immediately prior to the onset of cooling between 120 and 115 Ma.

 
Figure 3: Palaeotemperature constraints derived from AFTA data in individual samples in each well, plotted against depth. Summary stratigraphic columns are also shown for comparison (details in Figure 1). Palaeotemperature constraints are coded for discrete palaeo-thermal episodes as identified in Figure 2. Present day temperature profiles, together with corrected BHT data, are also shown for each well. Profiles parallel to the present-day temperature profiles are drawn through the palaeotemperatures characterising individual episodes in each well, as an aid to later discussion. Two sets of VR data are available from this well, with more reliable data represented by solid symbols.

Well 42/16-1: New results from Triassic and Carboniferous units in this well reveal at least three palaeo-thermal episodes which, on the basis of data from all four wells, are interpreted as representing the Early Cretaceous, Late Cretaceous-Early Tertiary and Late Tertiary episodes described earlier. Vitrinite reflectance values between 0.85 and ~1.2% from the Carboniferous section define maximum palaeotemperatures between 130 and 160°C. AFTA samples from similar depths were totally annealed prior to Late Cretaceous to Early Tertiary cooling, and provide a minimum estimate of 105°C in that episode. AFTA data from the overlying Triassic section were totally annealed prior to Early Cretaceous cooling, and provide a minimum estimate of 120°C in that episode, while providing discrete estimates of peak palaeotemperatures during two subsequent episodes. In the deeper AFTA sample, effects of the Early Cretaceous episode were overprinted by the Late Cretaceous – Early Tertiary episode. Thus, as with well 42/12-1, the maximum palaeotemperatures indicated by VR are slightly higher than the Early Cretaceous values revealed by AFTA, and this raises the question of whether the VR data represent this or an earlier episode (since the values from the two systems differ by only ~10°C, interpretation in terms of a common event would seem to be justified, but this cannot be definitely confirmed from these data alone). Once again, comparison of data from other wells (discussed later) strongly supports an interpretation in which the VR data do indeed represent the Early Cretaceous episode. As illustrated in Figure 3, palaeotemperature constraints derived from AFTA and VR data for the three episodes can be described by linear profiles with gradients similar to present-day values, although for the two most recent episodes, at least, a wide range of alternative interpretations would also be possible. A single VR value reported from the Triassic section falls below the profile drawn through the Early Cretaceous palaeotemperature constraints. If we are correct in attributing maximum palaeotemperatures throughout the Triassic and older section to the Early Cretaceous episode, then VR data from the Triassic section should also reflect that episode, in which case the palaeotemperature constraint from Triassic VR data should be co-linear with the profile through the deeper samples for this episode. Experience in the region suggests that VR data from the Triassic section are prone to serious errors, as suitable lithologies for analysis are rare, and often represent contaminant material (from cavings or some other source). Therefore little significance is attributed to this value in the interpretation of these data favoured here.

Well 42/17-1: Results from this well were also described in detail by Duncan et al (1998). As mentioned earlier, two AFTA samples collected from this well failed to yield any apatite. Thermal history information for this well is therefore dependent on VR data interpreted in a regional context and by comparison with data from neighbouring wells. VR values between 0.8 and ~1.3% from Carboniferous units suggest maximum palaeotemperatures between 120 and 160°C, while a single value from the Jurassic section gives a value around 80°C. As with the Triassic VR value in the 42/16-1 well (discussed above), this Jurassic value may be unreliable, and omitting this value, the data can be described by a linear profile with a slope similar to that of the present-day temperature profile. In the absence of AFTA data from this well, no direct indication is available of the timing of the maximum palaeotemperatures derived from the VR data. Comparison of values from this well with those in wells 42/12-1 and 42/16-1 (see later discussion) again strongly suggests that the VR data from this well represent the Early Cretaceous episode.

Well 42/21-1: New AFTA data from this well show a major difference to the three wells discussed so far, in that they clearly show that the thick Jurassic section intersected in this well began to cool from maximum palaeotemperatures in the Late Cretaceous to Early Tertiary (some time between 70 and 55 Ma), and show no evidence of Early Cretaceous effects. Late Tertiary cooling is also detected from AFTA, as illustrated in Figures 2 and 3. Maximum palaeotemperatures derived from VR data in this well are highly consistent with those from AFTA, as shown in Figure 3, confirming that the preserved section in this well cooled from maximum palaeotemperatures in the Late Cretaceous to Early Tertiary. However, one point to note from the results for this well shown in Figure 3 is the lack of scatter in the VR values about the trend with depth. Typical VR datasets show appreciably more scatter than the results in this well, and we therefore view these data with some suspicion (although as noted earlier, the values are highly consistent with the results from AFTA, and the overall values are considered to be broadly correct).

Comparison of palaeotemperature profiles in different wells

Figure 4 shows a comparison of palaeotemperature profiles characterising the three palaeo-thermal episodes in different wells. In the upper plot, palaeotemperature constraints interpreted as representing the Early Cretaceous episode emphasise the remarkable similarity in these values from the 42/16-1 and 42/17-1 wells, while values from well 42/12-1 appear to be slightly lower in magnitude while defining a similar overall trend. An interpretation of all these data in terms of a common episode offers the simplest explanation of all these results, and there seems no reason to invoke any earlier episodes to explain the VR data.

The central plot in Figure 4 emphasises the similar overall nature of Late Cretaceous-Early Tertiary palaeotemperatures in the three wells. Results from the 42/16-1 well appear to be higher than in the other two wells by around 15°C, but overall the similarity is most striking.

The lower plot in Figure 4 emphasises the similarity between Late Tertiary palaeotemperatures in the 42/16-1 and 42/21-1 wells. AFTA data from the 42/12-1 well do not show any evidence of the Late Tertiary cooling episode, as those data are dominated by the earlier episodes. But based on the similarity of data from the other wells, it seems likely that Late Tertiary cooling also affected the section preserved in this well, as well as that in the 42/17-1 well, in which no AFTA data are available and thus no constraints are possible on Late Tertiary effects.

Figure 4: Palaeotemperature constraints characterising the Early Cretaceous (upper), Late Cretaceous-Early Tertiary (centre) and Late Tertiary (lower) palaeo-thermal episodes in each well are plotted together on common axes, to facilitate comparison. For each palaeo-thermal episode, the magnitude of maximum or peak palaeotemperatures in different wells are remarkably similar.

Characterising mechanisms of heating and cooling

As illustrated in Figures 3 and 4, the variation of palaeotemperature with depth characterising the three palaeo-thermal episodes in all four wells can be described by linear profiles, more or less parallel to the present-day profile. We can assess this more quantitatively by applying the methods described by Bray et al. (1992) to define the range of palaeogeothermal gradient and removed section which are consistent (within 95% confidence limits) with the palaeotemperature constraints for each episode in each well.

Figure 5 shows the results of this procedure. For each palaeo-thermal episode, the range of allowed values of palaeo-gradient and removed section are shown, with results from individual wells compared on a common scale. The range of present-day gradients in these wells is also highlighted in this plot, and it is clear that results for all three palaeo-thermal episodes, while allowing broad ranges of both parameters in most cases, are consistent with an overall interpretation in which palaeogeothermal gradients have remained close to present-day values since at least the Early Cretaceous. It should be noted that in performing these analyses we have omitted the Jurassic VR value from well 42/17-1 and the Triassic VR value in 42/16-1. Inclusion of these data produces much less consistent interpretations, such that results from the 42/16-1 and 42/17-1 wells require much higher palaeogeothermal gradients that are not allowed by results from the 42/12-1 well. Given the close proximity of all these wells, a consistent explanation of results from all three wells seems much more likely, and therefore these Mesozoic VR data are excluded on the basis that they are unreliable.

Figure 5: These three plots show the range of values of palaeogeothermal gradient and removed section required to explain the palaeotemperature constraints characterising the Early Cretaceous (upper), Late Cretaceous-Early Tertiary (centre) and Late Tertiary (lower) palaeo-thermal episodes in each well, for a palaeo-surface temperature of 6°C. The effects of higher palaeo-surface temperatures can be allowed for as described in the text. Vertical shaded regions highlight the range of present-day thermal gradients in the four wells. Results from all four wells are consistent with a model in which palaeogeothermal gradients were close to present-day values since at least the Early Cretaceous, while significantly elevated palaeo-gradients appear to be ruled out for the Early Cretaceous episode in wells 42/12-1 and 42/17-1, and for the late Cretaceous-Early Tertiary episode in the 42/12-1 and 42/21-1 wells. Results for the Late Tertiary episode in wells 42/16-1 and 42/21-1 wells are consistent with a broad range of palaeo-gradients because of the narrow range of depths over which constraints are available (42/16-1) coupled with the quite broad range of palaeotemperatures allowed by the AFTA data in most cases (both wells).

Another point of note is that elevated palaeogeothermal gradients for the Late Cretaceous – Early Tertiary episode are clearly excluded by the palaeotemperature constraints provided by AFTA in the 42/12-1 and 42/21-1 wells, in which only values close to the lower end of the range of present-day values are allowed. Results from the 42/16-1 well allow a broader range of gradients, because of the relatively narrow depth range over which palaeotemperature constraints are available, combined with the deepest sample in that well providing only a minimum limit on the maximum palaeotemperature. Similar comments apply to the Late Tertiary results from this well and the 42/21-1 well.

In producing the allowed range shown in Figure 5 for the Late Cretaceous – Early Tertiary episode in the 42/21-1 well, only the AFTA-based palaeotemperature constraints have been used. As already noted, the VR data show much less scatter than expected in a well-behaved data-set, which suggests these data should be treated with some caution. The VR values as reported define a much lower palaeo-gradient (around 10°C/km) which would make results from this well inconsistent with those from other wells. Given the overall consistency of the palaeotemperature values characterising this episode in Figure 4, a similar interpretation in all wells is much more likely than local differences of this nature between individual wells.

In summary, AFTA and VR data from the four wells analysed for this study can be interpreted as representing the effects of three palaeo-thermal episodes as listed earlier, and the palaeotemperature constraints characterising each episode can be explained in terms of linear profiles with palaeogeothermal gradients close to present-day thermal gradients in the region.

This suggests, in turn, that the most likely explanation of these palaeo-thermal episodes is that heating was due almost solely due to deeper burial, with little or no contribution due to elevated basal heat flow, which would be manifested in these results by significantly higher palaeogeothermal gradients compared to present-day values.

On this basis, Figure 5 shows that for the Early Cretaceous episode, deeper burial by around 3 km of additional section, subsequently removed by progressive uplift and erosion since the Early Cretaceous, is required to explain the observed palaeotemperatures. For the Late Cretaceous-Early Tertiary episode, around 2 km of additional burial is required, while for the Late Tertiary episode the appropriate figure is around 1 km.

These values were derived assuming a constant palaeo-surface temperature through time, equal to the present-day temperature of 6°C. If the palaeo-surface temperature was higher in the past, then the quoted values of removed section can be easily converted to apply to other values of palaeo-surface temperature by subtracting or adding the difference in depth equivalent to the difference between this value and the new palaeo-surface temperature, for the appropriate palaeogeothermal gradient. For instance, if the palaeogeothermal gradient was 50°C/km and the palaeo-surface temperature was 10°C higher than the value assumed in this report, the estimated eroded section should be reduced by 200 metres. Different heating rates can be allowed for in similar fashion, with an order of magnitude change in heating rate equivalent to a 10°C change in palaeotemperature (palaeotemperatures increase for higher heating rates, and decrease for lower heating rates). For typical values, the assumed heating rate will not affect the shape or slope of the palaeo-temperature profile significantly.

Note that while results from wells 42/12-1, 42/16-1 and 42/17-1 show that at least 1 km of section must have been removed between the onset of Early Cretaceous and the subsequent Late Cretaceous-Early Tertiary palaeotemperature peak, the total amount of section removed in this interval (and thus the amount of subsequent re-burial during the Late Cretaceous and Early Tertiary) is not otherwise constrained. It is possible that all of the ~3 km of additional Triassic to Early Cretaceous section responsible for producing the Early Cretaceous palaeotemperatures could have been removed, followed by deposition of ~2 km of Late Cretaceous section. Alternatively, ~2 km of section might have been removed during Early Cretaceous uplift and erosion, after which another ~1 km of Late Cretaceous section was deposited to produce the required ~2 km of additional burial during the Late Cretaceous-Early Tertiary. A wide variety of alternative scenarios are also possible, within the overall constraint of maximum burial depths required to explain the palaeotemperature data. Similar comments apply to the interval between the Late Cretaceous-Early Tertiary and Late Tertiary episodes in all four wells. Integration of the results of this study with regional geological information on thicknesses of overburden preserved in regions not affected by the events under discussion is required to provide better definition of this aspect of the geological evolution of the region.

Thermal history synthesis

Reconstructed thermal histories for units intersected in Central Irish Sea Basin wells 42/16-1 and 42/21-1, based on the results presented in preceding sections, are shown in Figure 6. These reconstructions employ a constant palaeogeothermal gradient equal to the present-day values (30°C/km in well 42/16-1, 31.9°C/km in well 42/21-1), and the corresponding values of removed section from Figure 5. Given the overall similarity in results from the 42/12-1, 42/16-1 and 42/17-1 wells, the reconstructions in wells 42/12-1 and 42/17-1 are likely to be very similar to that illustrated for well 42/16-1.

In detail, to produce the reconstructions shown in Figure 6 for well 42/16-1 an additional 3475 metres of post-Upper Triassic sediment is deposited between 208 and 125 Ma, 2975 metres of which are removed by uplift and erosion between 120 and 110 Ma; a further 2050 metres are deposited between 110 Ma and 65 Ma; 2250 metres are removed by uplift and erosion between 65 and 60 Ma; a further 1000 metres are deposited between 60 Ma and 15 Ma; the remaining total of 1300 metres of additional section are removed between 15 Ma and 2 Ma.

For well 42/21-1, an additional 2000 metres of post-Oxfordian sediment is deposited between 155 and 65 Ma, with 1500 metres removed by uplift and erosion between 65 and 60 Ma, followed by a further 500 metres deposited between 60 Ma and 15 Ma, and the remaining total of 1000 metres of additional section removed between 15 Ma and 2 Ma.

 

Figure 6: Reconstructed thermal histories for units preserved in wells 42/16-1 and 42/21-1. Equivalent histories for the 42/12-1 and 42/17-1 wells are thought to be very similar to that shown here for well 42/16-1. While results from well 42/16-1 show no evidence of any end-Carboniferous (Variscan) palaeo-thermal effects, appreciable heating and cooling may have occurred at that time provided that peak palaeotemperatures were less than the maximum values reached in the Early Cretaceous episode. Similar comments apply to the possibility of Early Cretaceous heating and cooling in the 42/21-1 well.

Note that while the onset of cooling in these reconstructions is shown as 120 Ma, 65 Ma and 15 Ma in the three episodes, any time between 120 and 115 Ma, between 70 and 55 Ma, and between 25 and 0 Ma, respectively, would be allowed by the AFTA data from these wells.

Note also that as mentioned earlier, for well 42/16-1, the proportion of s of additional section deposited in the earliest episode which was actually removed prior to the recommencement of burial in the late Cretaceous is not defined precisely. In these reconstructions, we have selected one particular option out of many, simply to serve as an illustration of the overall nature of the history. Similar comments apply to the later episodes in both wells.

It should be appreciated that while we have assumed here that heating was solely due to deeper burial, alternative scenarios are possible and a wide variety of combinations of palaeo-gradient and removed section are capable of satisfying the palaeotemperature constraints from AFTA in these wells, as shown by the contoured regions in Figure 5. However, all such combinations of palaeogeothermal gradient and removed section in each well result in reconstructed thermal histories for the preserved units which are very similar to those shown in Figure 6, being tightly constrained by the AFTA and VR data presented in each well.

Parameters such as possible non-linearity of the palaeotemperature profiles, particularly through the removed section where no constraints are available, complicate the estimation of removed section, while palaeo-surface temperatures may well have been higher than the values assumed here, as noted earlier. In addition, heating rates may have differed from those assumed in obtaining the palaeotemperature constraints from AFTA and VR in these wells. All these factors may introduce systematic errors into the estimation of burial depths from the palaeotemperature data, and thus exact reconstruction of burial histories in these wells is difficult. But from the simple considerations outlined earlier, it is clear that these type of factors can account for only a few hundred metres of removed section, while higher palaeogeothermal gradients, which might reduce the amount of required burial, can be ruled out, as discussed earlier. Thus the palaeotemperature data clearly require removal of a total of around 3 km in well 42/16-1 and around 2 km in well 42/21-1, but attempts to determine more precisely the amounts of removed overburden using such approaches are not justified due to the various uncertainties involved.

However, we emphasise that despite these uncertainties in reconstructing former burial depths, the reconstructed thermal histories for units within the preserved section shown in Figure 6 are not subject to any of these uncertainties. Thus, the main aspects of the reconstructed thermal histories shown in Figure 6 are well constrained by the AFTA and VR data from these wells, and can be used with confidence to predict patterns of hydrocarbon generation etc.

Possible end-Carboniferous effects

In the thermal history reconstruction illustrated for well 42/16-1 in Figure 6, no significant deeper burial or palaeo-thermal effects are attributed to the Late Carboniferous to Early Triassic unconformity in this well. Given the regional occurrence of major heating and cooling associated with Variscan events represented by this unconformity, from onshore Ireland (see next section) across to South Wales and the North Devon coast, some palaeo-thermal effects undoubtedly affected the Carboniferous and older section in the CISB. But as discussed earlier, it seems clear that the Carboniferous section in wells 42/12-1, 42/16-1 and 42/17-1 reached their maximum palaeotemperatures during the Early Cretaceous. Nevertheless, it is worth noting that considerable additional burial and associated heating could have occurred during the time interval represented by the Variscan unconformity, provided that peak palaeotemperatures at this time did not exceed those reached in the Early Cretaceous episode.

Comparison with results from surrounding regions

One striking aspect of the results of this study is the similarity of the reconstructed thermal histories in three wells (42/12-1, 42/16-1 and 42/17-1), and in all four wells for post-Early Cretaceous time. The lack of detectable Early Cretaceous palaeo-thermal effects in the 42/21-1 well, stands in stark contrast to results from the other three wells.

Results from Central Irish Sea Basin wells 42/12-1, 42/16-1 and 42/17-1 are also highly consistent in most respects with those from neighbouring regions, where evidence for Early Cretaceous, Early Tertiary and Late Tertiary palaeo-thermal episodes is widespread. As noted at the beginning of this paper, and as illustrated in Figure 7, previous thermal history studies from the East Irish Sea Basin (Green et al., 1997) and from onshore Ireland (Green et al., 1998) have provided evidence of Early Cretaceous, Early Tertiary and Late Tertiary cooling episodes which correlate closely with those identified in the Central Irish Sea in this study. Reconstructed thermal histories for post Carboniferous times in these regions are almost identical to that illustrated for the 42/16-1 well in Figure 6. It seems reasonable to conclude that we are seeing the same episodes in all these areas, and thus these episodes appear to be truly regional in nature.

Figure 7: Reconstructed thermal histories from regions surrounding the CISB are remarkably similar to those identified in this study, at least for post-Carboniferous time. Histories for samples from onshore Ireland are taken from Green et al. (1998), those for the East Irish Sea Basin are from Green et al. (1997) and that for North Wales is from Duncan et al. (1998). Histories for South Wales are essentially identical to those shown for onshore Ireland (Geotrack, unpublished results). The palaeo-thermal episodes identified in the four CISB wells in this study appear to be of truly regional extent.

The main difference in the results from the CISB wells, compared to onshore Ireland and the EISB is the lack of detectable end-Carboniferous effects in the CISB. Reasons for this are unknown, but as discussed by Duncan et al. (1998), the preservation of Stephanian sediments in the CISB is consistent with the lack of pronounced Variscan erosion in this region.

Effects of Early Cretaceous cooling are also seen across Southern England (Bray et al., 1998), and SW England and SW Wales (unpublished Geotrack results), while Early Tertiary cooling is also recognised in these areas as well as Northern, Central and Eastern England (Green, 1989; Green et al., 1993; Lewis et al., 1992) and Northern Scotland (Thompson et al., 1999), further emphasising the regional extent of these episodes.

The dominance of individual episodes shows some variation, with the Early Tertiary episode for example being dominant in the East Irish Sea Basin (EISB), which has important ramifications for hydrocarbon prospectivity as discussed in the next section.

This is highlighted by the absence of detectable Early Cretaceous palaeo-thermal effects in well 42/21-1 in this study. This is thought to be due to the location of this well within a separate structural regime to the other three wells, aligned with the North Celtic Sea, St Georges Channel and Cardigan Bay Basins. As reported by Murdoch et al. (1995), AFTA data from the North Celtic Sea Basin shows that the Early Tertiary cooling episode is dominant in that region, and results from well 42/21-1 provide similar conclusions. Appreciable Early Cretaceous cooling may have affected these basins, but with peak palaeotemperatures less than subsequent maximum values reached in the Early Tertiary (analogous to the masking of postulated Variscan effects in well 42/16-1 in Figure 6).

Implications for regional hydrocarbon prospectivity

In the EISB and adjacent areas, Green et al., (1997) showed that the area in which source rocks reached maximum maturity levels immediately prior to Early Tertiary inversion is restricted largely to the main EISB hydrocarbon province. In surrounding areas, by contrast, the main phase of hydrocarbon generation occurred during earlier episodes (from end-Carboniferous to Early Cretaceous). The lack of hydrocarbon discoveries in these regions suggests either that the hydrocarbon generation pre-dated structure formation, or that any hydrocarbons accumulated in earlier episodes were lost during subsequent uplift and/or tilting.

Results presented here indicate that potential source rocks within most of the CISB reached maximum maturity levels during the Early Cretaceous, which represents the termination of the main phase of hydrocarbon generation. Any hydrocarbons accumulated at that time are likely to have undergone phase changes and redistribution during at least three discrete phases of uplift and erosion, significantly decreasing the chances of commercial amounts surviving to the present day.

To the south, in the vicinity of well 42/21-1, the Early Tertiary episode appears to become dominant, raising the possibility that in this region conditions similar to those characterising the main EISB hydrocarbon province may apply. However, due to the much larger thicknesses of preserved Jurassic section, any Carboniferous source rocks in most of this region are likely to have reached much higher maturity levels than in the EISB, while Jurassic source rocks are only marginally mature. These factors suggest much higher levels of exploration risk in this area, although the Dragon discovery in UK Quad 103 (Tanner, 1999) shows that conditions suitable for operation of a viable petroleum system existed in this region at some stage.

In conclusion, it is clear that significant additional risk is associated with timing of hydrocarbon generation in the CISB. In future exploration in the Central Irish Sea Basin and adjacent regions, this risk can be much reduced through recognition of the major palaeo-thermal episodes which have affected the region, and the variation in the magnitude of their effects across the region, in order to identify regions where the main phase of hydrocarbon generation post-dated structuring. As with the EISB, definition of areas in which the main phase of hydrocarbon generation occurred during the Early Tertiary or later is likely to highlight the most prospective areas.

Acknowledgments

We are grateful to PAD, Dublin for provision of sample material from four Central Irish Sea Basin wells for this study. AFTA® is the registered trademark of Geotrack International.

References

Armstrong, J.P., D’Elia, V.A.A. and Löberg, L. 1995. Holywell Shale: A potential source of hydrocarbons in the East Irish Sea. In: P.F. Croker and P.M. Shannon (Eds.), The Petroleum geology of Irelands Offshore Basins. Geological Society, London, Special Publications, 93, 37-38.

Armstrong, J.P., Smith, J., D’Elia, V.A.A. and Trueblood, S.P. 1997. The occurrence and correlation of oils and Namurian source rocks in the Liverpool Bay-North Wales area. In: N.S. Meadows, S. Trueblood, M. Hardman and G. Cowan (eds). Petroleum Geology of the Irish Sea and Adjacent Areas, Geological Society, London, Special Publications, 124, 195-211.

Bray, R., Green, P.F. & Duddy, I.R. 1992. Thermal history reconstruction in sedimentary basins using apatite fission track analysis and vitrinite reflectance: a case study from the east Midlands of England and the Southern North Sea. In: Hardman, R.F.P. (Ed) Exploration Britain: Into the next decade. Geological Society, London, Special Publications, 67, 3-25.

Bray, R., Green, P.F. and Duddy, I.R. 1998. Multiple heating episodes in the Wessex Basin: implications for geological evolution and hydrocarbon generation. In: J. R. Underhill (Ed.) Development, Evolution and Petroleum Geology of the Wessex Basin. Geological Society, London, Special Publications. 133, 199-213.

Burnham, A.K. and Sweeney, J.J., 1989. A chemical kinetic model of vitrinite reflectance maturation. Geochimica et Cosmochimica Acta., 53, 2649-2657.

Colter, V.S. 1997. The East Irish Sea Basin – from caterpillar to butterfly, a thirty year metamorphosis. In: N.S. Meadows, S. Trueblood, M. Hardman and G. Cowan (eds). Petroleum Geology of the Irish Sea and Adjacent Areas, Geological Society, London, Special Publications, 124, 1-9.

Corcoran, D. and Clayton, G. 1999. Interpretation of vitrinite reflectance profiles in the Central Irish Sea area – implication for the timing of organic maturation. Journal of Petroleum Geology (in press).

Duddy, I.R., Green, P.F. and Laslett G.M. 1988. Thermal annealing of fission tracks in apatite 3. Variable temperature behaviour. Chemical Geology (Isotope Geoscience Section), 73, 25-38.

Duddy, I.R., Green, P.F., Hegarty, K.A. and Bray. R.J. 1991. Reconstruction of thermal history in basin modelling using Apatite Fission Track Analysis: what is really possible. Proceedings of the First Offshore Australia Conference (Melbourne). III-49 - III-61.

Duddy, I.R., Green, P.F., Bray, R.J. and Hegarty, K.A. 1994. Recognition of the Thermal Effects of Fluid Flow in sedimentary Basins. In: Parnell, J. (ed.), 1994 Geofluids: Origin, Migration and Evolution of Fluids in Sedimentary Basins, Geological Society, London, Special Publications, 78, 325-345.

Duncan, W.I., Green, P.F. and Duddy, I.R. 1998. Source rock burial history and sea effectiveness: Key facets to understanding hydrocarbon exploration potential in the East and Central Irish Sea Basins. AAPG Bulletin, 82, 1401-1415.

Floodpage, J., White, J. and Newman, P. 1999. The hydrocarbon prospectivity of the Irish Sea: insights from recent exploration of the Central Irish Sea, Peel and Solway Basins (extended abstract). Petroleum Geology of Ireland’s Offshore Basins, Dublin, April 1999.

Green, P.F. 1989. Thermal and Tectonic history of the East Midlands shelf (onshore U.K.) and surrounding regions assessed by apatite fission track analysis. Journal of the Geological Society of London, 146, 755-773.

Green, P.F., Duddy, I.R., Gleadow, A.J.W., Tingate, P.R. and Laslett, G.M. 1986, Thermal annealing of fission tracks in apatite 1. A qualitative description. Chemical Geology (Isotope Geoscience Section), 59, 237-253.

Green, P.F., Duddy, I.R., Gleadow, A.J.W. and Lovering, J.F. 1989a. Apatite Fission Track Analysis as a palaeotemperature indicator for hydrocarbon exploration. In: Naeser, N.D. and McCulloh, T. (eds.) Thermal history of sedimentary basins - methods and case histories, Springer-Verlag, New York, 181-195.

Green, P.F., Duddy, I.R., Laslett, G.M., Hegarty, K.A., Gleadow, A.J.W. and Lovering, J.F. 1989b. Thermal annealing of fission tracks in apatite 4. Quantitative modelling techniques and extension to geological timescales. Chemical Geology (Isotope Geoscience Section), 79, 155-182

Green, P.F., Duddy, I.R., Bray, R.J. and Lewis, C.L.E. 1993. Elevated palaeotemperatures prior to early Tertiary cooling throughout the UK region: implications for hydrocarbon generation. In: Parker, J.R. (ed) Petroleum Geology of Northwest Europe: Proceedings of the 4th Conference, Barbican, London, 1067-1074.

Green, P.F., Duddy, I.R. and Bray, J.R. 1995. Applications of Thermal History Reconstruction in inverted basins. In: Buchanan, J.G. & Buchanan, P.G. (eds.), Basin Inversion, Geological Society, London, Special Publications, 88, 149-165.

Green, P.F., Hegarty, K.A. and Duddy, I.R., 1996. Compositional influences on fission track annealing in apatite and improvement in routine application of AFTA^®. AAPG, San Diego, Abstracts with Program, page A.56.

Green, P.F., Duddy, I.R. and Bray, R.J., 1997. Variation in thermal history styles around the Irish Sea and adjacent areas: implications for hydrocarbon occurrence and tectonic evolution. In: N.S. Meadows, S. Trueblood, M. Hardman and G. Cowan (eds). Petroleum Geology of the Irish Sea and Adjacent Areas, Geological Society, London, Special Publications, 124, 73-93.

Green, P.F., Duddy, I.R., Hegarty, K.A. & Bray, R.J. 1998. The post-Carboniferous history of Ireland: evidence from Thermal History Reconstruction. Conference Abstract, Ireland since the Carboniferous, Belfast, September, 1998.

Hardman, H, Buchanan, J, Herrington, P. and Carr, A., 1993. Geochemical Modelling of the East Irish Sea : its influence on predicting hydrocarbon type and quality. In: Parker, J.R. (ed.) Petroleum Geology of North West Europe: Proceedings of the 4th Conference. Geological Society, London, 809-821.

Laslett, G.M., Green, P.F., Duddy, I.R. and Gleadow, A.J.W. 1987. Thermal annealing of fission tracks in apatite 2. A quantitative analysis. Chemical Geology (Isotope Geoscience Section), 65, 1-13.

Lewis, C.L.E., Green, P.F., Carter, A. and Hurford, A.J. 1992. Elevated late Cretaceous to Early Tertiary palaeotemperatures throughout North-west England: three kilometres of Tertiary erosion? Earth and Planetary Science Letters, 112, 131-145.

Maddox, S.J., Blow, R. and Hardman, M. 1995. Hydrocarbon prospectivity of the Central Irish Sea Basin, with reference to Block 42/12. Offshore Ireland. In: P.F. Croker and P.M. Shannon (Eds.), The Petroleum geology of Irelands Offshore Basins. Geological Society, London, Special Publications, 93, 59-77.

Murdoch, L.M., Musgrove, F.W. and J.S. Perry 1995. Tertiary uplift and inversion history in the North Celtic Sea Basin and its influence on source rock maturity. In: P.F. Croker and P.M. Shannon (Eds.), The Petroleum geology of Irelands Offshore Basins. Geological Society, London, Special Publications, 93, 297-319.

Scotchman, I.C. and Thomas, R.W. 1995. Maturity and hydrocarbon generation in the Slyne Trough, northwest Ireland. In: P.F. Croker and P.M. Shannon (Eds.), The Petroleum geology of Irelands Offshore Basins. Geological Society, London, Special Publications, 93, 385-411.

Stuart, I.A. and Cowan, G. 1991. The south Morecambe Field, blocks 110/2a, 110/3a, 110/8a, UK East Irish Sea. In: Abbots, I.L. (ed.) United Kingdom Oil and Gas fields, 25 years commemorative volume. Geological Society, London, Memoir, 14, 527-541.

Summer N.S. and Verosub, K.L. 1989. A low temperature hydrothermal maturation mechanism for sedimentary basins associated with volcanic rocks, In: Price, P.A., (ed.) Origin and evolution of sedimentary basins and their economic potential. American Geophysical Union Geophysical Monograph, 48, 129-136.

Sweeney, J.J. and Burnham, A.K., 1990. Evaluation of a simple model of vitrinite reflectance based on chemical kinetics. AAPG Bulletin, 74, 1559-1570.

Tanner, H.C. The petroleum geology of the St Georges Channel Basin (extended abstract). Petroleum Geology of Ireland’s Offshore Basins, Dublin, April 1999.

Thompson, K., Underhill, J.R., Green, P.F., Bray, R.J. and Gibson, H.J. 1999. Evidence from apatite fission track analysis for the post-Devonian burial and exhumatuion history of the northern Highlands, Scotland. Marine and Petroleum Geology, 16, 27-39.

Ziagos, J.P. and Blackwell, D.D. 1986. A model for the transient temperature effect of horizontal fluid flow in geothermal systems. Journal of Volcanology and Geothermal Research, 27, 371-397.