Design of observation network
Observation targets of DONET are micro to large earthquakes, slow slip of the plate boundary seafloor deformation, and Tsunamis that occur beneath the area of DONET network. Currently, no long-term seafloor observatory exists in this area. Availability of no seafloor data is large limitation for us to observe these events in the seafloor. We aim, by DONET installation, not only to improve detection ability of events, but to model these events quantitatively with better accuracy.
Number of observatories for DONET installation is limited from technical, financial limitation. We designed our submarine cable observation system to increase the number of observatories significantly (in an order of ten) from present submarine cable network systems for earthquake monitoring which allow only several seismometers installed in each observation node. Increased number of observatory may be installed by extending cable from each observation node. Using multiple extension fiber-optic cables from observation node, our design allow maximum of 40 observatories from five observation nodes connected to backbone submarine cable. In the initial installation of DONET, by optimizing observatory density and location, we plan to install 20 observatories distributed densely, covering the area from the trench axis to the main rupture area of the last Tonankai earthquake. This distribution of observatory enables us to precisely determine hypocenter of small to large earthquakes and detect relatively small ground deformation in the seafloor. We evaluated this ability of DONET network by computer simulation.
These figures shows results from our simulation. Distribution of existing land seismic station (blue dots) and DONET observatories (blue triangles). Precision of magnitude 2 earthquake location simulated using only land observatories (Left) and both land and DONET observatories (Right), are shown. Hypocenter accuracy better than 5 km is shown by gray shaded area and epicenter accuracy better than 5 km is shown by red line. Area of reliable hypocenter resolution from only land observatory data (Left) is only in land. In the seafloor, it is clear that we cannot expect good accuracy for earthquake depth. Even when we look at epicenter location (not depth of earthquake), area of reliable location (shown by red line boundary) is limited in the seafloor. With DONET data, we can resolve both epicenter and depth of earthquake in the seafloor beneath the DONET network (Right).
By simulating with different DONET observatory distribution, we optimized our network so that area of good hypocenter location is maximized, yet there is no obvious gap of such area in and around the DONET network.
We also evaluated detection ability of ground deformation due to small precursory slip event. With precision observation of seafloor pressure change in densely distributed seafloor quartz pressure gauges, we expect to identify occurrence of small ground deformation. By such observation, slow slip event on the plate boundary may be detected. An event of moment magnitude 6 slip on the plate boundary in the seafloor was evaluated for detection by DONET network. Our simulation result suggests that such detection is possible if such event occurs below the DONET network, and seafloor pressure gauge has resolution of 1 cm or better. Such detection of events in the seafloor is impossible only with current GPS geodetic observation network which are dense, but only on land.
The observation targets of DONET, such as small to large earthquakes, slop-slip events on the plate boundary, Tsunamis, require precision seismometers and pressure gauges. A small earthquake has very small amplitude and typically observed in frequencies higher than 1 Hz. Large earthquakes show very large acceleration up to 2 G or so. On the other hand, slow slip events may be observed between 0.01-0.1 Hz, and even lower frequencies such as days or months. In all, dynamic range and frequency band of DONET observation target are wide (smaller than 10-9 m/s2 to 2 G) and broad (from 1/years to 100 Hz). To cover the dynamic range and frequency band, we plan to combine two types of seismometer and three types of pressure gauges installed in each observatory. Two types of seismometer are a broadband seismometer to cover weak motion in the frequency band of 1/360 Hz to 100 Hz and a strong motion accelerometer to cover strong motion in the frequency band from DC to 100 Hz. The three types of pressure gauges are quartz pressure gauge to observe Tsunamis and seafloor deformation by absolute pressure from DC to approximately 1 Hz, a differential pressure gauge to observe broadband seismic waves as a small change of seafloor pressure from 1/200 Hz to 20 Hz, and a hydrophone to observe high frequency acoustic waves. A differential pressure gauge is very sensitive and can detect pressure change of 0.1 Pa or smaller in frequencies 0.02-0.08 Hz, where very low frequency earthquakes are known to be observed.
To determine the seismometers to combine for the DONET observatory, we evaluated noise characteristics of seismometers of different kind in a vault of Matsushiro Seismological Observatory, Japan Meteorological Agency. Evaluated seismic sensors were three types of broadband seismographs (Streikeisen STS-2, Guralp CMG-3T, prototype version of Kinemetrics Cronos), three types of strong motion accelerometers (JAE JA-5 typeIII, JAE JA-40G, Metrozet TSA-100S), and four types of geophones.
In DONET network observation, we aim to observe slow slip event by seafloor pressure change. As well to monitor Tsunamis generated in and around the network, to monitor ground deformation in vertical direction, quartz pressure gauge, that gives depth of seafloor in terms of seafloor pressure, is used. Quartz pressure gauges are stable over years typically better than 0.5 psi/year. Effect of seafloor deformation due to slip events on the plate boundary is usually very small. Therefore, the stability of quartz pressure gauge used is very important, and have to be evaluated. We evaluated quartz pressure gauges for their long-term stability in environment similar to the seafloor. We installed quartz pressure gauges in a laboratory, simulating seafloor environment by dead weight gauge and water bath to maintain constant pressure at 4000 psi and constant temperature at 4°C. Manufacturer of tested quartz pressure gauges were Hewlett Packard, Paroscientific, and Clark Oilfield Measurement. Two different full-scale sensors, of 10000 psi full-scale and 6000 psi full-scale from Paroscientific quartz pressure gauge were evaluated. The laboratory test for more than continuous 150 days period was conducted to determine the type of quartz pressure gauge for the DONET observatory.
To isolate seismometers from the effect of seafloor current flow, we plan to design the seismometer package to enable surfacial burial in the sediment, while the pressure gauge package will be installed in the seafloor. From experience of previous seafloor deployments of Tsunami-meters, care will be taken to minimize temperature change of pressure gauges. These are the part of actions taken to achieve very low noise observation by the future DONET observatory.
We’ve compared and evaluated the background noise data between the buried type seismometer and the seafloor type seismometer at the Sea of KUMANO in 2007. To find out what kind of influence will occur to the data with different conditions (pressure, temperature or underflow). The results show that the buried type Seismometer with less background noise than the seafloor Seismometer at most condition. With the vertical motion, these two results were the same under a calm water condition. With the Horizontal motion, the buried Seismometer data showed about 10 times quieter than the seafloor Seismometer.
Buried type seismometer
Seafloor type seismomete