This essay has been submitted by a student. This is not an example of the work written by our professional essay writers.
Materials from 10 onshore North Island oil exploration wells have been examined by Cope Reed. They then proposed correlations of the indurated sandstone and schist in the wells with various facies of the New Zealand geosyncline (Eastern Province). Wodzicki (1974) examined basement material from four offshore oil exploration wells in the area of Fig. 2 and showed that various western province igneous and metamorphic rocks were represented. The provenance of Cretaceous-Cenozoic sandstones in the western and eastern Taranaki Basin broadly reflect a derivation from Western and Eastern Province sources, respectively (e.g., Smale 1992).
In this project, we present an interpretation of the distribution of geological units beneath Taranaki Basin. The interpretations are based on the:
Fig 1.1: Ma p of the Taranaki Basin region showing a seismic line where data was collected. The gas fields are in green and the field is in red.
Fig 1.2: Schematic Interpretation for the NM-16 Regional Seismic Line
Figure 1.3: Stratigraphic Framework of the Taranaki Basin (Adapted from King and Thrasher (1996))
Onland geological units are, of course, defined not just on the basis of their petrological content but also by using fossil, stratigraphic, lithofacies, and structural data. Despite the fact that the oil exploration well core and cuttings do not provide this extra information, we are confident that our petrological-geochronological approach yields valid interpretations.
1.1 GEOLOGICAL OVERVIEW:
Taranaki Basin (Fig. 1) is a major, Late Cretaceous-Cenozoic sedimentary basins in central New Zealand, that are, respectively, hydrocarbon-producing and hydrocarbon-prospective. The clastic sedimentary rocks of the basin were all ultimately derived from erosion of the underlying and adjacent pre-Late Cretaceous crystalline basement units of which are exposed in the North and South Islands. The distribution of geological units beneath and adjacent to these two basins has important implications for the Paleozoic and Mesozoic tectonic evolution of New Zealand.
The area straddles the North and South Islands, regions which are commonly treated separately in tectonic analyses (e.g., see comments by Black 1994). It is valuable to know if the recognised terranes and igneous suites continue between the two islands and to the north and west of New Zealand. Knowledge of the sub-basin geology also helps with studies of the provenance and paleo-geography of basin strata and in the reconstruction of reservoir sandstone depositional systems.
The onland basement rocks of New Zealand are well characterised on a regional (i.e. 1:1 000 000) scale and they can be most simply divided into Eastern and Western Provinces that are separated by the Median Tectonic Zone (MTZ). The Eastern Province is dominated by Late Paleozoic-Mesozoic indurated sandstones and mudstones with subordinate mafic volcanics and chert, in part overprinted by the Haast Schist. The Western Province consists of early Paleozoic siliciclastic and carbonate rock, intruded and metamorphosed by mid-Paleozoic and Cretaceous granitoids. The MTZ is characterised by a region of Carboniferous and Early Triassic to Early Cretaceous volcanic, plutonic, and sedimentary rocks, whose nature and contacts with the flanking Eastern and Western Provinces are the topic of ongoing research (Kimbrough et al. 1994). The rocks of the Eastern and Western Provinces have been divided into a number of petrographically and geochemically distinct tectonostratigraphic terranes and igneous suites (Fig. 2). Details of these divisions are beyond the scope of this paper, but recent summaries have been provided by Roser & Korsch (1988), Tulloch (1988),Bradshaw (1989), Cooper & Tulloch (1992), Mortimer (1993, 1995), Black (1994), Kimbrough et al. (1994), and Muiretal. (1994).
1.1.1 Taranaki Basin:
The detrital modes of sandstones from onshore Taranaki Basin wells Kiore-1 and Pukearuhe-1 are distinctly less quartz and rock fragments. The abundance of tuffaceous and calcareous material in the sandstones is also typical of Murihiku sandstones (e.g., Boles 1974) and atypical of other Eastern Province terranes. The geochemical composition of sandstones from the above five wells are permissibly Murihiku, but cannot be uniquely distinguished from Caples-Waipapa sandstones using chemical criteria alone (Fig. 4A).
Zeolites within the well further suggest a correlation with the Murihiku Terrane (cf. descriptions by Boles 1974; Black et al. 1993). The absence of prehnite-, pumpellyite-, or epidote-bearing assemblages rules out a correlation with Maitai, Caples, Waipapa, and Rakaia Torlesse Terranes.
1.1.2 Basin and Sub-Basin Structure
Interpretation of seismic reflection lines in the area by Thrasher & Cahill (1990), and Lewis et al. (1994) has identified a number of major faults that offset the basement-cover unconformity by up to 6 km. The Taranaki and Manaia Faults (Fig. 6) are two of the most important faults in this area and they both have had a long and complex history (e.g., King & Thrasher 1992). Mortimer (1993) has drawn attention to the fact that the Picton Fault (Fig. 6) is a major metamorphic and structural boundary within the Marlborough Schist Belt.
From experience in onshore South Island, the boundaries of Mesozoic terranes often coincide with Cenozoic faults. Given the lack of precision in locating any contacts between basement geological units beneath the Taranaki Basin, we have deliberately chosen to position them along these major basement-cover faults where geometry permits.
1.1.3 Processing of Seismic Data
Important steps were taken in the processing of the seismic data acquired to improve its signal quality and reduce its noise component. The main tasks of the seismic processing exercise are;
To correct for recording strategies and ray-path geometries.
To take advantage of what the recording techniques provide.
Enhance the signal to noise ratio (S/N)
To provide the clearest image of the subsurface.
The first step in the process is to perform a refraction muting exercise on the acquired data. This is done due to the fact that there is noise generated during the exercise which would be characterised as a direct arrival. The noise can also obliterate or overshadow the effect and power of the signal such that there is a reduction in the signal to noise ratio. The effect of the mute is to remove the noise generated and ensure that the signal is sampled for the most part.
Figure 1.4: Diagram showing the field record being muted to improve its signal to noise ratio, thereby increasing or improving the quality of the signal..
After this process, the data undergoes a quality control exercise, usually regarded as a QC. The effect of performing a data QC is to improve the quality of signal extracted from the field record by removing the effect of other sources of noise which may not be directly observable or traceable.
Figure1.5: Section of the field record showing shot gathers produced after noise reduction by Quality Control (QC). Notice the vertical lines close to each record which may have been produced from surface sources during the marine acquisition.
The data can then be analysed to obtain the frequency components within a certain time window. This is done to produce an estimate of the dominant frequency components within the data.
Figure 1.6: Graph showing the Frequency/ Power Spectrum of the Seismic Signal averaged over 120 traces within a time window of 0-3000 ms.
A frequency-wave number spectrum of the data after the quality control has been done can also be generated to produce an F-K spectrum of the data. This is expressed below;
Figure 1.7: Map of the F-K Spectrum of the Data after Quality Control
1.4 Noise in Seismic Data
Attenuating the high-amplitude noises, such as swell noise, seismic interference, ground roll and multipls, is really a big challenge in the seismic data processing. The final subsurface image may provide wrong information for the interpretation without attenuating these high-amplitude noises. Regarding the multiple attenuation, different techniques have been discussed in many literatures (Berkhout, 1982; Verschuur et al, 1992; Guo, 2003). We wouldn't discuss them in paper. The emphasis here is how to attenuate the swell noises, seismic interference noises and ground roll.
According to the characteristic of the high-amplitude noises, these high-amplitude noises can be divided into non-coherent high-amplitude noises, such as swell noises, highly dispersed ground roll, etc., and coherent high- amplitude noises, such as seismic interferences, ground-roll, etc.
Conventional way of attenuating the non-coherent high- amplitude noises is first to detect these high-amplitude noises, and to scale down or mute them. In this way, the noises are attenuated; however, unfortunately the signals are attenuated as well. Of course this is not what we expected.
Conventional way of attenuating the coherent high-amplitude noises is to apply a F-K filter or a pâˆ’Ï„filter to the data within a given range. The drawback with these algorithms is that these filters are deterministic, which introduces some limitations when signals and noises have overlaps in F-K or pâˆ’Ï„domain.
How to attenuate these kinds of noises has been a critical issue in the seismic data processing. An algorithm was introduced to attenuate high-amplitude noises by using a prediction filter to recover the frequency components contaminated by the high-amplitude noises(Guo,1993), and later another similar algorithm was introduced by using a projection filter(Soubaras,1995). In order to attenuate the non-coherent high-amplitude noises, we use a strategy that first detects the frequency components that are contaminated by the high-amplitude noises, and then these frequency components are replaced by filling in based on the frequency components of the neighbor traces using a projection filter. Using the strategy that first predicts the noises and then subtracts them from the input data can attenuate coherent high-amplitude noises. We use LSQR to automatically estimate the coherent high-amplitude noises, and then adaptively subtract them from the input data based on a pattern-based algorithm (Guo, 2003). Production implementations of these techniques show encouraging results.
High-amplitude noises include non-coherent high-amplitude noises and high-amplitude coherent noises. We use different strategies to attenuate them.
(a) Non-coherent high-amplitude noise attenuation
In the marine environments, swell noise is one of the non-coherent high-amplitude noises. It is one of the key problems needed to solve. Sometimes the swell noise is so strong that we couldn't identify the signals from the data. How to attenuate the swell noise has been one of our key efforts.
The frequency content of swell noise usually falls in the low frequency band. Abnormal frequency component, which is contaminated by the swell noises, can be detected within this low frequency band when the residual of a prediction error filter(PEF) exceeds a threshold, also the abnormal frequency component can be identified by comparing its amplitude with those of the neighbor traces. If its amplitude is greater than a threshold defined by the median value of the neighbor traces, it will be considered to be the abnormal value and should be replaced by a better frequency component. This better frequency component is obtained by filling in based on the frequency components of the neighbor traces using a projection filter. The relation between the projection filter and the PEF can be formulated using the matrix of the projection filters.
ï€¨bï€© Coherent high-amplitude noise attenuation
As described previously, coherent high-amplitude noises, such as seismic interferences, can be attenuated by the strategy that first estimates the noises and then adaptively subtracts them from the data.
Local linear events in f-x domain are spatially predictable. In most cases, coherent high-amplitude noises are locally linear, so this kind of noises is locally predictable. The strong coherent events can be estimated based a prediction error filter, which contains the information of the high- amplitude coherent noises. Based on these filters, the high- amplitude coherent noises can be obtained. These estimated high-amplitude coherent noises usually don't exactly match the real noises in the data, so next step is to adaptively subtract these events from the input data by using a matching filter.
Unfortunately, when signals strongly interfere with coherent noises, conventional adaptive subtraction algorithms give biased signals after adaptive subtraction. In order to solve the problem, a pattern-based algorithm is presented(Guo, 2003). We use this algorithm to subtract the coherent high-amplitude noises from the data. It can also be formulated using matrices.