T21C-0538 INVITED 0800h
Long-Term Observations of Active Hydrothermal Processes on the Gorda Ridge: The Sea Cliff Hydrothermal Field and Escanaba Trough
The two known sites of high temperature hydrothermal venting on the Gorda Ridge are the northerly Sea Cliff hydrothermal field and the southerly Escanaba Trough. Indications that the Sea Cliff field (GR-14) existed were first obtained from hydrographic work in 1985, and confirmed by the discovery of the site in 1988. Our cruise at the site in 2000 was the first time fluids were sampled, and we subsequently also collected fluids in 2002. One reason the Sea Cliff field is of interest is its location $\sim$3km east of the axis of spreading. It was suggested that the site might therefore be relatively 'old,' or at least located on 'older crust' at the 5.5cm/yr spreading rate. All of the hydrothermal fluids from Sea Cliff are low chlorinity, or vapor phase, contrary to its previous interpretation as an older hydrothermal system which are generally interpreted to vent fluids with chloride contents greater than seawater. Our other chemical data also support our interpretation that this system is not tapping highly altered crust. In spite of the $>$$300\deg$C measured temperatures of the vent fluids, they are strikingly clear, and have Fe contents $\sim30$ umoles/kg, about 2 orders of magnitude less than typical. The low Fe, and other transition metals, are likely a result of the slightly elevated pH (4.5 at $25\deg$C) of these fluids. Possible causes of the elevated pH include: incorporation of buried organic matter, dissolution of fracture-filling calcite, supercritical phase separation phenomena, and other reactions that may impact the proton balance in the fluids, as there is no sediment cover at this site, and no (chemical) evidence for buried sediments. The very low metal contents are in agreement with the water column signals observed in 1985, suggesting the fluids have not changed in $\sim20$ years, and that the site was not impacted by the seismic activity on the Gorda Ridge in 1996 and 2001. Hydrothermal fluids were first collected from Escanaba Trough (NESCA) in 1988, and this site was subsequently drilled by ODP in 1996. Our fluid samples collected in 2000 and 2002 have the same compositions as those collected in 1988. All of the fluids we collected venting from this site are high chlorinity (brines), while ODP also encountered low chlorinity vapors subsurface. We propose a model to explain why the brines are venting preferentially to the vapor phase at this site. Chemical stability in hydrothermal systems, as in these two cases from the Gorda Ridge, are in stark contrast to observations of variability elsewhere on the global mid-ocean ridge system. The time scales and explanations for the fundamental causes of these differences remain poorly understood.
T21C-0539 0800h
Long-Term Continuous Monitoring of Fluid Chemistry and Flux at the Bush Hill Gas Hydrate Field, Gulf of Mexico Using a New Flow Meter, The MOSQUITO
Long-term monitoring of fluid, solute, and methane fluxes that influence marine gas hydrate formation and dissociation has important implications for the seafloor biochemical environment, ocean chemistry, and potentially the atmosphere. Four newly designed flux meters called the MOSQUITO (Multiple Orifice Sampler and Quantitative Injection Tracer Observer) and two temperature loggers were deployed adjacent to the Bush Hill hydrate mound in the northern Gulf of Mexico (GC185) in order to understand how chemistry, physics, biology, and subsurface hydrology dynamically influence the growth and dissociation of the hydrate mound. The MOSQUITO contains a network of osmotic samplers and a tracer injection device, each connected to a titanium capillary tube that penetrates the sediment. The tracer is injected as a point source, and fluid chemistry and tracer concentrations are continuously sampled simultaneously at multiple depths below the seafloor in a three dimensional array with respect to the tracer injection point. Bottom water chemistry is also sampled continuously. Vertical and horizontal flow rates as low as 1 cm/yr are determined by modeling the variability in tracer concentration at each depth over time. MOSQUITOs can be deployed at passive margins, ridge crests, ridge flanks, subduction zones, and lakes. MOSQUITOs were deployed over a period of 430 days from June 2002 to August 2003 and were sampled at weekly resolution. The temperature loggers were attached to the MOSQUITOs and recorded seafloor temperature every 40 minutes. Three MOSQUITOs were deployed within 3 m of the hydrate mound and $\sim$ 5 m apart, adjacent to transient methane seeps; in a mussel field, in a bacterial mat, and in a tubeworm field. The fourth MOSQUITO was placed $\sim$150 m southwest of the hydrate mound to monitor background fluid flow, geochemistry, and temperature. The average bottom water temperature over the 430-day deployment period was $7.94\deg$C, with minimum temperatures occurring every 6-9 weeks and maximum temperatures occurring every 9-12 weeks. The record shows an asymmetric periodicity with a slower rise in temperature followed by relatively abrupt cooling. The hydrology around the hydrate mound is complex with both upflow and downflow of fluid occurring within each of the sub-environments. Within the mussel field, fluid upflow ranges from 1-39 cm/yr and downflow varies from 3-65 cm/yr, with the polarity of fluid flow changing at 1-7 week intervals. In the bacterial mats, upward flow velocities range from 1-100 cm/yr and downward flow velocities range from 4-130 cm/yr. The polarity of fluid flow in the bacterial mats appears to be linked to bottom water temperatures and is seasonal and less variable, with fluid downflow occurring from June to November 2002, as well as during the summer of 2003, and fluid upflow from November 2002 to mid-June 2003. The background site is characterized by minor upward fluid advection. Pore fluid chemistry also varies significantly over short distances and time, and is manifested at the seafloor by sharp interfaces between different benthic biological communities. The relations between bottom water temperature, pore fluid chemical fluctuations, and hydrology are being evaluated. Net chemical fluxes of the major elements and methane, as well as net fluid fluxes, and their relation to gas hydrate formation and dissociation, ocean chemistry, and biology will be addressed.
T21C-0540 0800h
Seismicity rates of slow, intermediate, and fast spreading ridges: Insights from long-term hydroacoustic monitoring
Ocean basin earthquakes recorded on NOAA/OSU and U.S. Navy hydrophone arrays are used to evaluate long-term volcano-tectonic seismicity levels from segments of the fast-spreading rate East Pacific Rise (EPR) from 20$\deg$S-20$\deg$N, intermediate-spreading rate Juan de Fuca Ridge (JdFR) from 39$\deg$-52$\deg$N and Galapagos Rift (GR) from 90$\deg$-103$\deg$W, and the slow-spreading northern Mid-Atlantic Ridge (MAR) from 5$\deg$-60$\deg$N. The hydrophones record the acoustic energy of seafloor earthquakes that propagate along the ocean sound channel with little attenuation over large distances. Frequency-magnitude relationships (Bohnenstiehl et al., 2002; Dziak et al., 2004) indicate the hydrophone catalogs are complete in these regions to body-wave magnitude $\sim$2.5 (EPR and GR), 2.5 (JdFR), and 3.0 (MAR), an improvement of 1.5 to 2 units over the land-based seismic catalogs for mid-ocean ridge systems. Using the hydrophone earthquake catalog, we will compare seismicity rates of the JdFR (12 years of data), to seismicity rates along the GR (6 years) and EPR (6 years) and MAR (4 years of data from 5$\deg$-39$\deg$ N; 16 months from 39$\deg$-60$\deg$ N). During these monitoring periods, five confirmed seafloor spreading events (four of which were associated with magmatic activity) were recorded on discrete JdFR segments, while 6 possible magmatic events were observed on the EPR, one on the GR, and one on the MAR. Empirical orthogonal functions will be used to elucidate the space-time patterns of seismicity and compare between the various spreading rates ridges, as well as to investigate the recurrence rate of seafloor spreading events present. In addition, single-link cluster analysis (SLC; Frolich and Davis, 1990) will be used to de-cluster the earthquake databases to reduce the effects of aftershock sequences and magmatic swarms, allowing us to evaluate how overall plate motion and changes in spreading rate effect levels of seismicity between ridge segments and different ridge systems. Preliminary results indicate the distribution of seismicity at the JdFR "super"-segments (between transforms) are positively skewed, indicating there are significant, yet brief periods of time (1-2 months) when seismic activity is well above the mean. Transforms along the JdFR, however, exhibit a random distribution.
T21C-0541 0800h
Earthquake Source Parameters Inferred from T-Wave Observations
The seismicity of the North Atlantic Ocean has been recorded by two networks of autonomous hydrophones moored within the SOFAR channel on the flanks of the Mid-Atlantic Ridge (MAR). In February 1999, a consortium of U.S. investigators (NSF and NOAA) deployed a 6-element hydrophone array for long-term monitoring of MAR seismicity between 15$^\circ$-35$^\circ$N south of the Azores. In May 2002, an international collaboration of French, Portuguese, and U.S. researchers deployed a 6-element hydrophone array north of the Azores Plateau from 40$^\circ$-50$^\circ$N. The northern network (referred to as SIRENA) was recovered in September 2003. The low attenuation properties of the SOFAR channel for earthquake T-wave propagation results in a detection threshold reduction from a magnitude completeness level (Mc) of $\sim$ 4.7 for MAR events recorded by the land-based seismic networks to Mc=3.0 using hydrophone arrays. Detailed focal depth and mechanism information, however, remain elusive due to the complexities of seismo-acoustic propagation paths. Nonetheless, recent analyses (Dziak, 2001; Park and Odom, 2001) indicate fault parameter information is contained within the T-wave signal packet. We investigate this relationship further by comparing an earthquake's T-wave duration and acoustic energy to seismic magnitude (NEIC) and radiation pattern (for events M$>$5) from the Harvard moment-tensor catalog. First results show earthquake energy is well represented by the acoustic energy of the T-waves, however T-wave codas are significantly influenced by acoustic propagation effects and do not allow a direct determination of the seismic magnitude of the earthquakes. Second, there appears to be a correlation between T-wave acoustic energy, azimuth from earthquake source to the hydrophone, and the radiation pattern of the earthquake's SH waves. These preliminary results indicate there is a relationship between the T-wave observations and earthquake source parameters, allowing for additional insights into T-wave propagation.
T21C-0542 0800h
Long-Term Seismicity of Northern (15$^\deg$N-60$^\deg$N) Mid-Atlantic Ridge (MAR) Recorded by two Regional Hydrophone Arrays: a Widespread Along-Ridge Influence of the Azores and Iceland Hotspots
The seismicity of the North Atlantic was recorded by two networks of hydrophones moored in the SOFAR channel, north and south of the Azores Plateau. The interpretation of the hydro-acoustic signals recorded during the first six-month common period of operation of the two networks (June 2002 to Nov. 2002) provides a unique data set on the spatial and time distributions of the numerous low-magnitude earthquakes which occurred along the Mid-Atlantic Ridge. Close to 2000 events were localized during this six-month period between latitudes 15$^\deg$N and 63$^\deg$N, 501 of which are localized within the SIRENA network (40$^\deg$N-51$^\deg$N) and 692 within the wider South Azores network (17$^\deg$N-33$^\deg$N). Using hydrophones to locate seafloor earthquakes by interpreting T-wave signals lowers the detection threshold of Mid-Atlantic Ridge events to 3.0 mb from the 4.7 mb of global seismic networks. This represents an average thirty-fold increase in the number of events: 62 events were detected by global seismological networks within the same area during the same period. An along-ridge spatial distribution of the seismicity is obtained by computing the cumulated numbers of events in 1$^\deg$-wide latitudinal bins. When plotted vs. latitude, this first-order distribution shows remarkable long-wavelength patterns: the seismicity rate is low when approaching the Azores and Iceland (reaching values as low as 10 events/d$^\deg$), while it peaks to 70 events/d$^\deg$ in the vicinity of the Gibbs FZ. Moreover, the latitudinal distribution of the seismicity hints at an asymmetric influence of the Azores hotpot on the MAR. Finally, the spatial distribution of the seismicity anti-correlates well at long wavelengths with the zero-age depths along the MAR and correlates with the zero-age Mantle Bouguer (MBA) anomaly values and the Vs velocity anomalies at 100 km in the upper mantle. It is thus proposed that the seismicity level would be partly tied to the rheology and thickness of the brittle layer and be thus dependant on the thermal regime of the upper mantle. The seismicity distribution could then be used as an additional tool to characterize the along-ridge influence of the Azores and Iceland hotspots on the MAR slow-spreading center.
T21C-0543 0800h
Long-term Spatial and Temporal Variations in Seismic Activity Along the Juan de Fuca Plate System (Northeast Pacific Ocean) Recorded on Military Hydrophone Arrays
Since August 1991 U.S. Navy hydrophones have been used by NOAA/OSU to monitor seismicity from seafloor spreading centers in the northeast Pacific Ocean. Over the last decade, more than 10,000 earthquakes have been located using the Tertiary (T-) wave of these seafloor events recorded by this hydro-acoustic monitoring system. Before a reliable analysis of the long-term behavior of northeast Pacific seismicity could be undertaken, a magnitude completeness filter (~208 dB) as well as a single link cluster (SLC) algorithm (Frolich and Davis, 1990) were applied to each segment of the ridge and plate systems. The SLC technique reduces clusters of earthquakes caused by large tectonic sequences or volcanic eruptions to single events in the catalog. This removes any bias due to large events from the monthly average of seismicity over the record length. From analysis of the long-term monthly average of events, a substantial increase in seismicity is observed in the entire northeast Pacific earthquake database beginning in 1998 with a peak in late 2000 and trailing back down during 2001. Events per month during this period of heightened seismicity often double the mean value ($\mu$ = 43.8 $\pm$ 2.0 events/month). The entire record was also broken down into the geologic segments of the Explorer Ridge, Sovanco Transform, Juan de Fuca Ridge, Blanco Transform, and Gorda Ridge revealing similar increases in seismicity along each segment contributing to the overall increase observed in the composite time series. First order statistics reveal the Blanco ($\mu$$_{Bl}$ = 20.1 $\pm$ 0.9 events/month) as the most active component followed by the Gorda ($\mu$$_{Gr}$ = 8.3 $\pm$ 0.5 events/month) and Juan de Fuca segments ($\mu$$_{Jf}$ = 6.7 $\pm$ 0.7 events/month). In addition, the distribution of seismicity at ridge segments are positively skewed, indicating there are significant, yet brief periods of time when seismic activity is well above the mean whereas transforms exhibit a more random distribution. Empirical orthogonal functions (EOF) will be used to further decompose the time-space modes of variability in seismicity both within and between ridge and transform segments, as well as to investigate the recurrence rate of seafloor spreading events present in the non-declustered earthquake database. We will also explore correlations of bottom pressure with long-term signals in seismicity as well as any correlation with the slow earthquakes at the Cascadia Subduction Zone documented by {\it Szeliga et al., 2004}.
T21C-0544 0800h
Standing-acoustic waves observed at the 2003 Tokaci-oki earthquake
At the 2003 Tokachi-oki, Japan earthquake, we have obtained the first actual dataset of phenomena such as sea-floor uplift and responding sea-water mass movement using a cabled observatory installed on the sea-floor (Watanabe et al., 2004). Among them, the acquired pressure waveforms at two pressure gauges (PGs) for the epicentral distances of 31.4 km and 81.8 km show significant fluctuations with respect to both the amplitudes and the periods during the earthquake. The PGs have recorded pressure fluctuations with peak-to-peak amplitude of approximately 4000 hPa and 1000 hPa on PG1 and PG2, respectively, and the dominant periods of 6 s have been recorded on both PGs. After the earthquake, the pressure waveforms have left significant static changes in the sea-floor, having 40 hPa and 10 hPa at PG1 and PG2, respectively. Note that 1 hPa change in static water pressure roughly corresponds to 1 cm of the ocean-bottom motion. The amplitude of dynamic pressure is approximately 100 times as large as that of static one. By using this unique instrument, we have found that the sea-floor uplift could reproduce the high amplitude acoustic wave in addition to the tsunami at the time of the earthquake. We introduce that tsunami and acoustic waves can be controlled by static displacement and velocity of the sea-floor under the assumption of constant velocity uplift. The acoustic wave bounces up and down between the sea-floor and the sea-surface, and it forms a standing wave. Since both the tsunami and acoustic waves are caused by the same phenomena, one constrains the other to obtain real sea-floor uplift in a time series. The order estimate of the acoustic wave_fs amplitude led us that the abrupt change in water depth has taken place in a time duration of several seconds. Also, our study reinforces that acoustic waves might have a potential use for the rapid tsunami warning. We think that the observation of the M8 earthquake on the seafloor has broadened the extent of tsunami studies from static to dynamic seafloor movement at the time of earthquakes.
T21C-0545 0800h
Long Term Seismic Observation in Mariana by OBSs : Results of the DD inversion
In order to obtain the deep arc structural image of Mariana, a large-scale seismic observation by using 58 long-term ocean bottom seismometers (LTOBS) had been performed from June 2003 until April 2004, which is a part of the MARGINS program funded by the NSF. Prior to this observation, a pilot long-term seismic array observation was conducted in the same area by using 10 LTOBSs from Oct. 2001 until Feb. 2003. By using seven LTOBS's data, those are about 11 months long, hypocenter determination was performed at first and more than 3000 local events were found, although the PDE list contains only 59. A 1D velocity structure based on the iasp91 model was used, and a systematic shift of epicenters between the PDE list and this study was seen. To investigate the detail of hypocenter distribution and the 3D velocity structure, the DD inversion (tomoDD: Zhang and Thurber, 2003) was applied for this data set with a 1D structure initial model except for the crust, which has been surveyed by using a dense airgun-OBS system (Takahashi et al., 2003). The result of relocated hypocenters shows double seismic zones until about 200km depth and a lined focuses along the current ridge axis in the back-arc basin, and the result of the tomographic inversion shows a image of subducting slab and a low-Vs region below the Pagan island erupted in 1981 at 80km depth. The mantle structure beneath the back-arc basin was not clearly resolved due to the inadequate source-receiver coverage, which is cleared in the recent experiment.
T21C-0546 0800h
First Results From the Deployment of a Buried Broadband Seismometer on the Endeavour Segment of the Juan de Fuca Ridge
The W.M. Keck Foundation with additional support from the Monterey Bay Aquarium Research Institute (MBARI) and the University of Washington is supporting a five-year program to conduct prototype seafloor observatory experiments to constrain the relationships between episodic geological deformation, fluid venting and microbial productivity along the boundaries of the northern Juan de Fuca Plate. The seismic component of the program comprises two types of instruments: ten short-period corehole seismometers deployed in local seismic networks on the Endeavour segment of the Juan de Fuca Ridge and at the intersection of the Nootka fault and the Cascadia subduction zone; and three broadband seismometers distributed regionally, one in each local network and the third on the Explorer plate. The broadband instruments are Guralp CMG-1T three-axis seismometers that are sensitive over a frequency range of 2.8 mHz (360 sec) to 50 Hz. To reduce the effects of water currents on the recorded data, each sensor is completely buried in sediment inside a 60 cm deep by 60 cm diameter caisson. A 20 m cable connects the sensor to an MBARI/GEOSense LP1 data logger and a 30 kW-hr battery sitting on the sediment. The first broadband was deployed as part of the Endeavour local network in the summer of 2003 and it represents the first buried broadband seismometer to be deployed on a mid-ocean ridge. In August 2004, we successfully recovered the first year of data from this instrument and deployed the remaining broadband seismometers. An initial inspection of the Endeavour broadband data reveals many examples of teleseismic, regional, and local earthquakes. For teleseismic earthquakes, the signal-to-noise ratio in the 0.025-0.1 Hz frequency band appears comparable to stations on Vancouver Island. We will present a preliminary analysis of the broadband data including an assessment of noise levels and data quality.
T21C-0547 0800h
Electrical Conductivity Structure Beneath the Northwest Pacific Basin Revealed by Long-term Electromagnetic Observation at the Seafloor
A seafloor geoelectromagnetic observatory has been operated for more than three years since 1st of August, 2001. The site, NWP, is located at (41 07'03"N, 159 55'43"E, 5570m) in the northwest Pacific, which is over 1300 km away from any of existing land geomagnetic observatories. The seafloor observatory is measuring temporal variations of the Earth_fs electromagnetic field with one-minute interval in addition to very precise attitude data to recover the same reference frame as in land observatories. NWP was chosen not only for seaward extension of the land geomagnetic observatory network to increase our knowledge of the Earth_fs geomagnetic field in time and space, but also for looking deep into the Earth beneath the seafloor as old as 120 Ma. The long and simultaneous time-series of the seafloor electric and magnetic fields measured at NWP allow us to delineate the electrical conductivity structure all through the upper mantle using magnetotellurics. It is expected that the oceanic lithosphere just before subduction is well-developed, whose electrical structure may fall in an end member model of oceanic plates. This information is indispensable to understand the whole cycle of the seafloor process from its birth to subduction. The asthenospheric structure beneath the very thick lithosphere is also important in terms of small-scale convection and water content, both of which are key issues of present mantle dynamics. It is also noteworthy that Hirano et al. (2001) reported a very fresh (6 Ma) basalt sampled at the seaward edge of Japan Trench as old as 130 Ma. To understand this new type of intraplate volcanism, it is important to construct a standard model of the electrical conductivity structure in the northwest Pacific that constitutes the counterpart of Lizarralde et al's (1995) model in the northeast Pacific.