Previous   |    Next

doi:10.2204/iodp.proc.332.105.2011

Results and discussion

Operations

The two SmartPlug instruments built in 2009 were set to begin recording data at the time they were shipped from the Pacific Geoscience Centre, Canada, to Shingu Port, Japan, on 11 April 2009, and stored in Shingu until they were transported to the D/V Chikyu via supply boat for Expedition 319. Because one of them (Instrument 82) was not deployed during the cruise, it was shipped back to Shingu and stored in a warehouse until 2010. It was loaded onto the Chikyu during the preexpedition port call on 26 October 2010 and hence contained an 18 month record of data by the time Expedition 332 began. Logging intervals for the formation and hydrostatic pressure sensors (P_form and P_hyd) and the internal platinum thermometer were set to 1 min; at this rate and with other operational parameters as set, battery power (provided by six Tadiran TL-5137 DD primary lithium cells) is the limiting factor for operational lifetime, which is roughly 7 y, including a de-rating factor of 75% applied to full power withdrawal. The instruments are equipped with 512 MB (low power) flash memory cards, which provide a (theoretical) storage capacity until the year 2038 at a 1 min sampling rate. The independent MTL in Instrument 82 was set to sample temperature at 60 min intervals. The main logger clock was synchronized to Universal Time Coordinated (UTC) on 11 April 2009, and the MTL clocks were set on approximately the same date. The clock was synchronized to UTC again on 6 November 2010 prior to deployment.

A third SmartPlug was fabricated in 2010 and followed an identical design. This instrument package was set to a sampling interval of 30 s, resulting in a (theoretical) storage capacity until the year 2033. Also, the software developed for programming and communication with the instrument was updated for the 2010 instruments. The third instrument was prepared so that it could be used as a backup in case time limitations on operations in Hole C0002J would not allow deployment of the planned long-term borehole monitoring system (LTBMS) (see “Long-term borehole monitoring system” in the “Methods” chapter [Expedition 332 Scientists, 2011]). All configurations and potential damage during shipping were checked prior to deployment onboard the Chikyu (Fig. F3).

The two GeniusPlug extension units were assembled during the first weeks of Expedition 332. As described above, the OsmoSamplers were designed for a deployment duration of up to 24 months. The FLOCS chambers were filled with various types of rock chips, including clayey silt recovered adjacent to the splay fault at Site C0004 (see Tobin et al., 2009, for details). All chambers and tubing were carefully saturated with sterile seawater so that no air got trapped inside before all components were placed in the extension unit at the bottom of the plug. As a final step, the tubing from the downward-looking pressure transducer was connected at the bullnose end cap using a Swagelok fitting (Fig. F5C). The first SmartPlug was directly attached to the bridge plug and was installed successfully during Expedition 319 on 23 August 2009 (Figs. F5A, F6A) (Expedition 319 Scientists, 2010). It was recovered on 7 November 2010 during Expedition 332 (Kopf et al., 2010). The GeniusPlug was mounted to a second bridge plug and deployed on 11 November 2010.

Results

Complete records are shown for the downward-looking (megasplay fault zone [P_form]; Fig. F7A) and upward-looking (hydrostatic reference [P_hyd]; Fig. F7B) pressure response and the pressure anomaly (i.e., excess pore pressure; Fig. F7C). The full records clearly show an increase in formation pressure and a subtle decrease in the seafloor reference but an overall increase in excess pore pressure.

The P records can be separated into six intervals. Period I, lasting until Day 61, shows P_hyd increasing at a rate of 1500 kPa/day, whereas P_form is falling at a similar rate. Period II (Days 61–154) is characterized by reference and borehole pressures that show a steady decline of 300 kPa/day and a steady increase of 360 kPa/day, resulting in excess pore pressure rises of 0.07 kPa/day. During Period III (Days 154–215), the rate of P decay (seafloor) and rise (formation) decreased to 35 kPa/day and 145 kPa/day, respectively. Period IV (Days 215–300) shows higher rates once again (P_hyd = –179 kPa/day, P_form = 291 kPa/day), resulting in the excess pore pressure rising at 0.05 kPa/day (Fig. F7C). The following Period V (Days 300–352) is characterized by rates of 22 kPa/day for P_hyd and 110 kPa/day for P_form. In contrast to other periods, pressure records show no opposite behavior but a simultaneous increase. Period VI (Days 352–452) shows a decreasing rate in P_hyd (269 kPa/day) and a rise to 430 kPa/day for P_form. Hence, the excess pore pressure shows an increase of 0.07 kPa/day (Fig. F7C).The records suggest that the borehole had not fully equilibrated by the time the instrument was recovered some 15 months after deployment. The data also demonstrate that the bridge plug effectively sealed the borehole because upon reentry of the drill string and latching onto the device during Expedition 332, the upward-looking pressure sensor shows a strong fluctuation owing to displacement of borehole fluid, whereas the downward-looking pressure sensor encounters no such interference and remains at a near-constant value (Fig. F8). Hydraulic isolation of the monitoring interval is also documented by (1) the observation that the ambient pressures are different by ~10 kPa for the two transducers and (2) the attenuation of the ocean tidal loading signal in the P_form record.

Regardless of the overall trend, three different types of transient pressure excursions were observed:

1. Pressure pulses of up to several 100 Pa in sudden amplitude change that lasted for 20–60 min, which are largely associated with teleseismic events;

2. Tremorlike signals of a fraction of a kilopascal with durations of hours or even days, which can be associated with atmospheric and oceanographic events such as low-pressure systems, storm waves, and tsunamis; and

3. Smaller individual pulses of low amplitude (<1 kPa) that lasted for only minutes, which are tentatively associated with either low-magnitude deformational events in the accretionary complex or the overlying seafloor.

In this short note, we only give one example for each of the phenomena mentioned above. However, postcruise research is ongoing and a detailed study will be published in the near future (e.g., S. Hammerschmidt et al., unpubl. data). Figure F9 shows the pore pressure anomaly associated with the Maule M 8.8 earthquake in Chile, which occurred on 27 February 2010 and was one of the most prominent subduction thrust ruptures in recent times (Moreno et al., 2010). Pressure waves produced 200 Pa in excess of P_hyd in the formation and arrived almost 24 h earlier than the tremorlike signal of the associated tsunami wave. The latter showed stronger P_hyd amplitudes compared to P_form, but maxima ranged only 60–75 Pa (see Fig. F9C). It can be seen that the incoming tsunami waves have periods of up to 10 min and that the signals span a period >12 h.

A second type of tremorlike signal, equally spanning over many hours or even days, is attributed to the transient pressure changes associated with low-pressure weather systems and storm waves and higher swell. Among the many examples that can be found during the period of SmartPlug monitoring, Typhoon Chaba is shown in Figure F10A. This low-pressure front hit the coast of Japan at the time the Chikyu was bound for Expedition 332 at Shingu Port, and in fact caused an early departure and a period of waiting on weather during late October 2010. The record attests that both P_hyd and P_form are loaded in a similar manner and magnitude.

The third type of short-period tremor, or sometimes just a series of individual pulses, is believed to result from internal deformation within the Nankai Trough accretionary complex. Those events of microseismic activity at very low frequencies were first described for the NanTroSEIZE area by Obara and Ito (2005) and attest that interseismic strain is not confined to slow elastic strain accumulation. Figure F10B shows a sudden ~350 Pa increase as a function of such a very low frequency event. Similar events were identified in various parts of the 450+ day record; however, two factors hamper their clear interpretation: (1) the relatively long sampling interval of 60 s, which may result in shorter deformation events being missed; and (2) the wealth of Japanese earthquake monitoring stations that have yet to analyze all of the data.

The temperature data are less meaningful than the P transients. Figure F11 provides an example of the full T data set comprising the two Paroscientific transducers, the platinum chip inside the pressure housing, and the self-contained ANTARES MTL, which show minor shifts but otherwise plot along the same narrow corridor. The data set attests that T is still recovering from the sudden drop during deployment, so a period >15 months is required to reach ambient background T values.

One peculiarity in the otherwise fairly homogeneous records from the four thermistors is an event around the 295th day into the monitoring phase. This sudden change in T lasted ~1.5 days and shows an increase in the MTL temperature as well as the T sensors in the two pressure transducers, whereas the sensor in the pressure housing experienced a drop in temperature. So far, no correlation with either the SmartPlug pressure data or with the United Stated Geological Survey or local Japanese earthquake (EQ) catalogs could be made.

Preliminary interpretation of the data set

Although detailed analysis of the full data set is ongoing, there are several key preliminary conclusions from the initial deployment and data review:

1. The Baker Hughes bridge plug effectively sealed the cased Hole C0010A as demonstrated by the significantly different response of the upward- and downward-looking pressure transducers when the drill string reentered Hole C0010A during SmartPlug recovery (see Fig. F8), the attenuated response of the formation pore pressure to tidal loading, and the ~10 kPa difference in ambient pressure recorded by the upward-looking and downward-looking pressure sensors.

2. It appears that the plug settled ~50 cm over the first 2 months of deployment or some of the thread grease or anticorrosion additive to the borehole fluid was mucking the inside of the instrument (and hence the upward-looking pressure transducer) and apparently loaded the seafloor sensor, resulting in a gradual increase in measured reference pressure of ~1500 kPa/day over this time interval (see Period I, Fig. F7).

3. The formation is slightly overpressured (~10 kPa) and appears to be still recovering at the end of the record (Fig. F7). The temperature record over the entire monitoring period supports this conclusion, with T still rising at the end of the record when the SmartPlug was recovered (Fig. F11).

4. The temperature records are well resolved but show yet-to-be-explained offsets, which may have something to do with a change in the thermal insulation around the sensors or the instrument pressure case, the changing nature of heat dissipation, or hydrogeologic events such as earthquake loading of the formation. Postcruise research may shed light on the observed problem of opposite temperature trends from different sensors during the same event (Fig. F11, inset).

5. Even with low-frequency sampling (60 s interval) the pressure records document responses to several earthquake events and associated Rayleigh and tsunami waves. The tsunami waveform related to the Maule, Chile, M 8.8 earthquake (see Fig. F9) is similar to existing records from, for example, the NEPTUNE seafloor cabled network (E. Davis, pers. comm., 2010) but has larger amplitude and a more persistent pressure signal lasting 3–4 days. Further, it demonstrates the more effective, frequency-dependent loading of the formation (seen in pressure fluctuations) from Rayleigh waves compared with tsunami loading.

6. The loading efficiency at the tsunami frequency is virtually identical to that of 0.82 at tidal frequency, illustrating the undrained, instantaneous response to overburden hydrostatic stress transients.

7. Subtle changes in pressure may be related to strain events within the Nankai accretionary prism, the most prominent of which occurred on 28 September 2009 (see Fig. F10B) and 5 January 2010. However, local EQ catalogs will need to be consulted in a future detailed study in order to confidently correlate pressure transients with tectonic strain events.

Top of page   |    Previous   |    Next