The GNSS buoy system for early tsunami warnings has been under development for about 20 years. A small prototype buoy was first deployed in Sagami Bay, Japan, in 1997. Then, after a series of experiments aiming for operational use, the system was implemented as a part of national wave monitoring system NOWPHAS. The NOWPHAS system had set up more than 10 GNSS buoys around Japan by 2011, and it recorded the tsunami caused by the 11 March 2011 Tohoku-oki earthquake. The records were used to update the tsunami warning at the time of the 2011 Tohoku-oki earthquake. However, the buoys were placed less than 20 km from the coast, as the system used the baseline mode RTK-GPS algorithm, which was not far enough for effective evacuation of people. Thus, we began trying to improve the system by putting the buoys much farther from the coast. The new system employs a newly developed positioning algorithm, Precise Point Positioning with Ambiguity Resolution (PPP-AR), together with satellite data transmission. A series of experiments involving the new system successfully indicated changes in sea level with an accuracy of a few centimeters. Given the success of the experiments, we are trying to use the GNSS buoys not only to provide early tsunami warnings but also to monitor various other geohazards. For example, we are trying to use the GNSS-Acoustic system to continuously monitor crustal movements on the ocean floor, to monitor the ionosphere, and to monitor the atmosphere. Ancillary sensors on the buoys will be utilized for oceanographic monitoring as well.

Using Global Navigation Satellite System–Acoustic (GNSS-A) technique, we have been developing observation system on a moored buoy for continuous monitoring of seafloor crustal deformation. The sound speed structure near a warm current has heterogeneity, which is the main cause of a seafloor positioning error. Assuming a sloping structure, previous studies proposed sound speed model to reduce positioning error. We examined the validity of the model by comparing the estimated structure with the actual structure measured at multiple points around our observation site. The result shows that the gradient parameter estimated from GNSS-A data acquired by vessel is appropriate. The numerical examination indicates that modeling error caused by the misinterpretation of the depth of gradient layer occurs, and it can be suppressed by performing acoustic ranging at the point near the centroid of units. From the calculation of estimation error of sound speed variation, the predicted acoustic ranging error observed using the moored buoy staying near the centroid is 9.0 cm or below. Therefore, seafloor displacement can be detected with centimeter class via moored buoy in the basin of a warm current.

In this paper, we describe the development of a continuous real-time system capable of measuring seafloor crustal deformation using the global satellite navigation system (GNSS)/Acoustic technique and a moored buoy. A program developed was implemented on the buoy to automatically distinguish the onset of a direct acoustic wave even if that wave had been contaminated with reflected waves and to detect the true travel-times by onboard processing rather transferring raw waveforms to the ground base station. This onboard procedure contributed to reduce the data size over a satellite communication. We conducted an operations test for a total of 106 days and found that the acoustic ranging and data transmissions were frequently interrupted by an unstable power supply, resulting in only 21% of the transmitted data being received at the ground base station. Nevertheless, we did not find any problem with continuous acoustic ranging measurement except for the above-mentioned power supply failure.