Core Slabbing and Nannofossil Analysis on the Chelungpu Fault Zone, Taichung, Taiwan

1 Exploration and Development Research Institute, Chinese Petroleum Corporation, Miaoli, Taiwan, ROC 2 Department of Earth Sciences and Institute of Geophysics, National Central University, Chung-Li, Taiwan, ROC 3 Department of Earth Science, National Taiwan Normal University, Taipei, Taiwan, ROC 4 Institute of Applied Geophysics, National Taiwan Ocean University, Keelung, Taiwan, ROC * Corresponding author address: Dr. Jong-Chang Wu, Exploration and Development Research Institute, Chinese Petroleum Corporation, Miaoli, Taiwan, ROC; E-mail: 048704@cpc.com.tw doi: 10.3319/TAO.2007.18.2.295(TCDP) The results of this nannofossil analysis supply essential information for determining the formation boundaries in the upper Pliocene to Pleistocene. These results also verify the existence of a repetition fossil zone. The TCDP well-A was sunk through the soft fine-grain muddy sandstone and mudstone dominated formations of the Pliocene and Pleistocene in the Taichung area. This study determines methods for providing core preservation in wells at fault zones and establishes a nannofossil biostratigraphy for the integrated Taiwan Chelungpu-fault Drilling Project (TCDP). Good core fabrics are useful for core description and sampling. In this present study, over 400 meters of subsurface cores were covered in resin and slabbed. Digitized images were created for all the core fabrics. More than 150 rock samples were analyzed for nannofossils to give a detailed appraisal of the biostratigraphic column of TCDP well-A. A fossil zone at a depth interval of 431 869 m is a NN16 18 biozone. This zone is within the Cholan Formation, a lithologic stratigraphy in northern and central Taiwan. The depth interval 883 1226 m is NN15, and is Terr. Atmos. Ocean. Sci., Vol. 18, No. 2, June 2007 296 within the Chinshui Shale. The Chelungpu fault is composed of five major shear zones. These are all found at depth within the marine Chinshui Shale. At a depth interval of 1293.37 1710 m is a NN12 14 biozone; this interval is within the Kueichulin Formation. Interestingly, both the interval beneath 1714 m and the nannofossil zone near the well bottom are NN16 18 (Cholan Formation), indicating a repeat of the Cholan Formation. The lowest fossil zone is also abundant in secondary reworked fossils in its assemblages. Hence, the repetition of the younger fossil zone, NN16 18, at the bottom of the well verifies the subsurface position of the Sanyi Fault and indicates that TCDP well-A must have passed through it. (

within the Chinshui Shale. The Chelungpu fault is composed of five major shear zones. These are all found at depth within the marine Chinshui Shale. At a depth interval of 1293.37 -1710 m is a NN12 -14 biozone; this interval is within the Kueichulin Formation.
Interestingly, both the interval beneath 1714 m and the nannofossil zone near the well bottom are NN16 -18 (Cholan Formation), indicating a repeat of the Cholan Formation. The lowest fossil zone is also abundant in secondary reworked fossils in its assemblages. Hence, the repetition of the younger fossil zone, NN16 -18, at the bottom of the well verifies the subsurface position of the Sanyi Fault and indicates that TCDP well-A must have passed through it.

INTRODUCTION
The Taiwan Chelungpu-fault Drilling Project (TCDP) is an integrated international study of the kinematics and dynamic behaviors of the 1999 Chi-Chi earthquake. TCDP well-A retrieved cores from 430 m to a depth of 2000 m (Figs. 1, 2). During the TCDP, preparation of the precious soft cores for different regimes became an important task. This project required lineate core preservation and abundant core sampling to satisfy the many core analyses that were to be conducted. Therefore, it was vital that all geologists communicated and cooperated well to their mutual benefit. The Exploration and Development Research Institute (EDRI) of the Chinese Petroleum Corporation (CPC) took over the task of core preservation and core slabbing using an internationally recognized standard process. It was then necessary to use low temperature storage to prevent temperature variation induced fractures in the cores. As expected, TCDP well-A passed through the Pleistocene and Pliocene half-soft rock formations of the Cholan, Chinshui Shale, and Kueichulin Formations. This fragile fault-related rock provided enhanced difficulties during core slabbing. However, the staff at the core laboratory in EDRI had some previous experience in processing this kind of half-soft rock in 2002 Hung et al. 2001).
These Plio-Pleistocene cores are classified as the upper sequences of the Taichung Basin, and are dominated by very fine-grain muddy sandstones and even mudstone facies. Therefore, age verification during drilling is necessary to preserve the integrity of the nannofossil record. The staff of EDRI also took over the task of stratigraphic analysis to help in the engineering progress and assist in identification of formations during well drilling.

CORE SLABBING
The procedures followed for the half-soft core slabbing of core samples for TCDP well-A are shown in Fig. 3. The first longitudinal cores retrieved were put into a cool room for storage. While the cores were dry, resin was applied to the upper surface of each core. After about 10 days, resin was applied to the remaining surface of the same cores. Then, these cores were cut into two symmetric portions. If, however, advanced secondary core slabbing was requested, this slabbing process was more complicated. Before a secondary cut could be made, it was necessary to resin the fresh surface of the first cut and then wait an additional 10 days for the core to dry. Finally, these core portions were divided into three parts: portion A to be kept for longterm preservation or post completion of a detailed description be available for limited sampling; portion B, the intermediate unit, to be permanently preserved and fixed to the bottom of a core box using epoxy resin; and portion C to be made available for analysis. Generally, the supply of portion C is not enough for most of laboratories so that it is necessary to resin the fresh core surface to protect the sample from being crushed during the slabbing process.   A typical example of core slabbing and a core photograph are shown in Fig. 4. Core photographs show detail of the lithology, sedimentary, and structural fabrics of the cores at different depth intervals. After pictures were taken of the entire core, close-ups were taken of the upper, middle, and lower parts of the core with overlap to allow for correlation in the digital database. The fabric in the detailed close-ups is much clear than that given by the core scanner.

NANNOFOSSIL ANALYSIS
TCDP well-A penetrates mudstone dominated strata making it difficult to discern the boundaries among the three formations. This affects the casing project and electrical logging plan during drilling. In this study, more than 150 samples were analyzed for nannofossil stratigraphy. In some cases the nature of the stratgraphy was immediately recognizable during the drilling process.

Methods
Samples were taken from the marine strata whenever possible; i.e., from shale, fine grain sandstone, or mudstone. After removing any surface material, a small amount of fresh sample (>30 g) was bagged, numbered, and logged for depth. When confronted with geological intervals that lacked conformity, it was important that more sample points were utilized and records were taken describing the unconformity of the contact features between the overlying and underlying strata.

Thin Section Preparation for Nannofossil Analysis
In the microfossil laboratory of EDRI (CPC), Dr. Gartner's method was utilized; however, Dr. Haq's method was kept for simple and easy processing in addition to avoiding the production of poisonous gas during heating of the Caedax resin. Dr. Gartner's method is highly suitable for the concentration of samples but is limited only to calcareous nannofossils.

Identification
After finishing rock-sample preparation, identification and counting of the assembled number was accomplished using high resolution Zeiss microscopes with polarized Nicolas. The identification criteria were based on the nannofossil zonation and definition developed by Martini (1971).

Fossil Photographs
The following fossil photographs were taken using a digital camera and prepared with CoreDraw software.

Fossil Statistics
The degree of richness in the distribution of nannofossils is completed using an xy dispersive pattern from an Excel applied format. Examining the richness distribution of the nannofossils helps to determine stratigraphic boundaries within each well column.

Illustration of Nannofossil Zones
Based on the classification of Martini (1971) (Fig. 5), Okada and Bukry (1980), and Chi et al. (1981), there are three nannofossil zones within the Chelungpu Fault area. At present, possible fossil zones encountered in TCDP well-A are described as follows:

NN12: Ceratolithus Acutus Zone
Definition: The first appearance of Ceratolithus acutus Gartner and Bukry as the lower boundary, the first appearance of Ceartolithus rugosus Bukry & Bramtette as the upper boundary, and all intermediate parts belong to NN12. The upper boundary determination is aided by the first appearance of Gephyrocapsa sp.

NN19: Pseudoemiliania Lacunosa Zone
Definition: The first appearance of Gephyrocapsa oceanica Kamptner as the lower boundary, the extinction surface of Pseudoemiliania lacunosa as the upper boundary, and all intermediate parts belong to this zone. The NN19 nannofossil zone is from Pleistocene. Correlation: In the NN19 zone, the top of Cholan Formation correlates to the Tokoushan Formation in the Hsinchu, Miaoli, and Taichung areas.

Results of Nannofossil Analysis
In reference to: the appearance of index fossils, the amount of nannofossils, fossil assemblages, the amount of secondary reworked fossils, lithology, and superposition of overlying and underlying strata, the 431.34 -2003 m geological column of TCDP well-A can be subdivided into 12 intervals. Each of these fossil assemblages will be described in the following way (Tables 1 -10).

Depth interval 431.34 -869 m:
This interval contains few fossils with no index fossils of Gephyrcapsa oceanica Kamptner found. The Gephyrcapsa oceanica Kamptner is very common in formations of the Pleistocene to Holocene; the non-existence of this fossil means that the strata are older than the Pleistocene. Therefore, it is necessary to refer to the underlying Bramlette (1969a); Reticulofenstra pseudoumbilica Gartner (1969c); Reticulofenstra minutula Haq & Berggren (1978); Discoaster pentaradiatus Bramlette & Riedel (1954); and Helicosphaera sellii Bukry & Bramlette (1969b). The Gephyrocapsa sp. occupies most parts of the interval while the Reticulofenstra minutula Haq & Berggren (1978) occupies the second part, and almost no Pseudoumbilica lacunosa Gartner (1969c) exist. This interval belongs to the NN15 fossil zone and is equivalent to early stage Pliocene.  Fert (1954) in the NN13 -15 zone; and Gephyrocapsa oceanica Kamptner in zone NN19 were not found. Therefore, it is deduced that the nannofossil biostratigraphy between 1712.7 and 2003 m correlates to the upper Pliocene (Figs. 5,11,12).
In conclusion, the nannofossil biostratigraphy between the well top and 869 m corresponds to zones NN16 -18, making this interval equivalent to the Cholan Formation. 882.6 m corresponds to the NN15 biozone and correlates to the Chinshui Shale. The interval 1111 -1313 m is the contact interval with the major Chelungpu fault zone and beyond this point to a depth of 1225.7 m is still within in the NN15 biozone, equivalent to Chinshui Shale. The interval 1292.37 -1710 m is judged to be within the NN12-NN14 biozones, correlating to the Kueichulin Formation. Below 1710 m, the well is thought to have penetrated the Sanyi Fault (Fig. 6) on account of the repeated appearance of dominate secondary reworked fossils. This interval is classified as NN16 -18, correlating to the Cholan Formation (Figs. 5, 6).

DISCUSSION
Since Pengli orogeny began, Neogene rocks in the western basin have been deformed folding and thrusting. At the same time, reworked fossils were deposited in the younger strata.
In the sedimentary strata of the Taiwan area, the index fossils of NN12 -14 are very rare       and difficult to subdivide, even though they exist in central and northwestern Taiwan (Table 11). Generally, to assist in making biostratigraphic analysis for the CPC exploration, the boundary between NN14 and 15 is used as criteria for both the sharp extinction of the large fossil Reticulofenestra pseudoumbilica and the sharp decrease of Sphenolithus abies. But through out the study area Reticulofenestra pseudoumbilica (large type) tends to be migratable and the distribution of Sphenolithus abies is not very uniform in central and northern Taiwan, making it difficult to accurately distinguish the boundary between NN14 and 15. Therefore, determination of the stratigraphic boundary between the Kuichulin and Chinshui Shales requires lithological analysis data. In order to constrain the nannofossil index for the subsurface fossil analysis of TCDP well-A, Table 11. Stratigraphy in central and northern Taiwan (Revised from Chi 1981;Huang 1982;Wu et al. 2002).
9 samples were taken from an outcrop near this well (Fig. 14). The analysis results are shown in Table 12, where sample no.1 is rare among fossils but whose fossil assemblage is possibly of the Pliocene age. Samples nos. 2 and 3 are NN15, equivalent to the Chinshui Shale; Sample no. 4 is NN12 -13, equivalent to the lower part of the Kueichulin Formation; Sample no. 5 is NN13 -15, equivalent to the upper part of the Chinshui Shale; Samples nos. 6 -8 are NN15, equivalent to the Chinshui Shale; and Sample no. 9 is NN16 -18, equivalent to the Cholan Formation. All the results of nannofossil analysis correlate strongly with the geological map of the CPC. Additionally, in a previous study, related samples from BH1, BH1A, and CLF-2 boreholes fall in the biozone range of NN14 -15 (Matini's fossil zone, Martini 1971), which is upper Early Pliocene in age and equivalent to the Yutenping Sandstone Member and the Chinshui Shale along the Holung-Chi sections of the Miaoli area, northern Taiwan. The boundary between the Chinshui Shale and the Kueichulin Formation in BH-1 borehole is located at 210.5 -225.6 m. This assessment is based on the distribution richness of nannofossil peak variation of the nannofossil content at the BH-1A borehole and probably originates from the change of parasequence between different delta systems within the Yutenping Sandstone. The boundary between the Chinshui and the Kueichulin Formations is judged by lithofacies and sedimentary facies . Nannofossil columns from these nearby wells should provide good correlation to those of TCDP well-A.
If we put the nannofossil zones, richness distributions, litho-stratigraphic, and structural features together (Fig. 13) we find that the Chinshui shale corresponds with the richest nannofossil zone, while the lower Cholan Formation is subordinate and the upper Cholan and Kueichulin Formations seem to lack large amounts of nannofossils. The Chinshui Shale is characterized by mudstone, some siltstone and rare fine-grain sandstone. The sedimentary features are dominated by aggradation rather than the progradation in gamma ray and lithologic columns. The paleo-environment of the Chinshui Shale is probably open marine. The overlying Cholan formation is dominated by very fine-grain muddy sandstone and interbed with siltstone and mudstone. Prograding features are highly developed in the lower part of this Cholan Formation (Lin et al. 2007). The underlying formation of the Chinshui Shale is the Kueichulin Formation, which is dominated by the fine-grain clean sandstone interbedded with very fine-grain sandstone and rare siltstone. The clean and coarse-grain sandstone occurs in the upper part of the Kueichulin Formation rather than the lower part. In the depth interval 1710 -2003 m of TCDP well-A, the nannofossil zone is determined to be NN16 -18 or part of the Cholan Formation, which is dominated by very fine-grain sandstone and interbedded with fine-grain sandstone and rare siltstone in lithology. The repetition of the Cholan Formation also determines the depth of the Sanyi Fault. The deep repeating section of the Cholan Formation differs slightly from the shallow Cholan Formation in lithology. Based on the structural interpretation of TCDP well-A and its geological section by Hung et al. (2007), the first drilled section of Cholan is the typical lower Cholan Formation, while the second drilled section is typical of the middle Cholan Formation. Our nannofossil analysis has proposed good criteria to help determine the Sanyi Fault zone within the muddy fine and very fine grain regime. Many Chelungpu Fault zones occur in the Chinshui Shale and the boundary between the Chinshui Shale and the Kueichulin Formation. These Chelungpu shear zones are actually the   (Lin et al. 2007) and structural  interpretations. Many Chelungpu Fault zones occur within Chinshui Shale (NN15) and the boundary between the Chinshui Shale and Kueichulin Formation (between NN15 and NN12 -14). Therefore, the Chelungpu Fault zone is better termed a no flexure bedding slip fault and it is difficult to find the fossil zone repetition while the Sanyi Fault determined by the bedding change and the clear repetition of the fossil zone. be the Sanyi Fault. Nonetheless, these nannofossils assemblages appearing in the interval 1717.7 -2003 m are very similar to those reworked assemblages in the Cholan Formation of the shallower part of the drilled well. Since in-situ fossil assemblages are very rare in the Cholan Formation, the characteristics of reworked assemblages is the only criteria for us to identify the appearance of the formation. Therefore, we suggest that the interval be the Cholan Formation and that the top of the repeated formation could be the depth of the penetrating Sanyi fault.

CONCLUSIONS
The TCDP well-A penetrates the fine-grain muddy sandstone and sandy mudstone formations of the Pliocene and Pleistocene in the Taichung area. The soft cores have difficulty satisfying all the requests of different geo science regimes. Therefore, core preservation by standard procedures and nannofossil analysis to determine formation boundaries has become very important, especially during well drilling. In this study, a 456 meter subsurface core has been coated in resin and slabbed, and core fabrics digitized. More than 150 rock samples have been analyzed for nannofossil to constrain a detailed biostratigraphic column on TCDP well-A. The fossil zones at depth intervals are given as follows: 431 -883 m is NN16 -18; 883 -1226 m is NN12 -15; 1239 -1710 m is NN12 -14, and the residual bottom interval is NN16 -18; an interval which is also abundant with reworked fossil assemblages. This study also reveals the repetition of the younger fossil zone NN16 -18 in the bottom interval of the well and verifies the exact subsurface position of the Sanyi Fault.