The Global Navigation Satellite System (GNSS) has been utilized in a variety of research fields within the geosciences. This research has been further developed for application to hazard monitoring and natural disaster mitigation. Some developments have even been implemented in society in countermeasures against natural disasters. The Geospatial Information Authority of Japan (GSI), for example, has established a nationwide GNSS network called GEONET. The data from GEONET are used extensively among researchers and practitioners, not only for basic research but also for the development of methods and systems that can mitigate disasters. This special volume is a collection of articles that discuss how such methods and systems are now being developed and/or planned to both clarify the mechanisms behind natural hazards and mitigate the damage they may cause. The volume consists of 13 papers covering a wide range of natural phenomena, such as earthquakes, crustal movements, tsunamis, ionospheric disturbances, and volcanic eruptions. Some papers help us to understand how natural hazards behave, which should be the first step toward disaster mitigation. On the other hand, other articles report direct efforts made toward providing early warnings of impending disasters. Disaster mitigation systems may require real-time (and even kinematic with high-rate data sampling) processing and dissemination of data. Moreover, some applications involve data collection from coastal waters and the open sea. Now that the density of GNSS stations has approached saturation on land, the scarcity of data collected offshore will have to be rectified through the development of GNSS systems in the ocean. We do hope that this volume will be a step in the further progress of utilizing GNSS for disaster monitoring and mitigation in the future to make society safer and more secure.
The Geospatial Information Authority of Japan operates one of the largest GNSS Continuously Operating Reference Station networks in the world, GEONET, consisting of more than 1,300 GNSS-based control points in Japan. GEONET has become a social infrastructure for surveying and mapping, monitoring crustal deformation, and mitigating natural disasters. It is also used for precise positioning to guide construction and farm machines, and even for weather forecasting. Here we review its history and outlook, focusing on its application to disaster mitigation. As Japan is surrounded by four tectonic plates with interactions that lead to frequent earthquakes and volcanic activities, it must maintain such a geodetic infrastructure for monitoring crustal deformations in as near to real-time for possible disaster mitigation.
Currently, the Geospatial Information Authority of Japan routinely operates the GNSS analysis system to produce coordinate solutions using a GNSS station network called GEONET. The sampling rate of these coordinate solutions is 1 day, or 3 hours at the highest. To augment the system to produce coordinate solutions with higher temporal resolution, a GNSS analysis system to produce kinematic coordinate solutions using the GEONET stations has been developed. This analysis system adopts the precise point positioning strategy to make the system robust and lightweight. In fact, the system is so lightweight that it can be operated with two personal computers. The system produces three types of coordinate solutions with intervals of 30 s, depending on their latency. Each coordinate solution has a quality represented by a typical coordinate repeatability of 1 cm in horizontal components. The system has been successfully utilized in constructing a fault model of the 2016 Central Tottori Earthquake. In the future, the system will be updated to host capacities for 1) producing post-processing 1 s coordinate solutions and 2) producing real-time 1 s coordinate solutions.
A new real-time Global Navigation Satellite System (GNSS) analysis system named REGARD has been launched to provide finite-fault models for large earthquakes with magnitudes =8 in real time. The finite-fault estimates using GNSS positioning are free from saturation problems and are very robust for modeling large earthquakes. The REGARD system processes ∼1,200 stations of GEONET, and event detection and finite-fault model inversion routines are implemented. Tests for the case of the 2011 Tohoku earthquake (Mw9.0) and a simulated Nankai Trough earthquake (Mw8.7) show that the REGARD system can provide reliable finite-fault models for large earthquakes. Furthermore, operational real-time results for the 2016 Kumamoto earthquake (Mj7.3) demonstrated the capability of this system to model inland earthquakes. These results imply the possibility of improving tsunami simulations and/or hazard information using rapid finite-fault models. Efforts to integrate real-time GNSS with current warning systems are currently being implemented around the world, and the REGARD system will join these systems in the near future.
This short paper reviews the role of real-time global navigation satellite system (GNSS) in near-field tsunami forecasting. Recent efforts highlight that coseismic fault model estimation based on real-time GNSS has contributed substantially to our understanding of large magnitude earthquakes and their fault expansions. We briefly introduce the history of use of real-time GNSS processing in the rapid estimation of the coseismic finite fault model. Additionally, we discuss our recent trials on the estimation of quasi real-time tsunami inundation based on real-time GNSS data. Obtained results clearly suggest the effectiveness of real-time GNSS for tsunami inundation estimation as the GNSS can capture fault expansion and its slip amount in a relatively accurate manner within a short time period. We also discuss the future prospects of using real-time GNSS data for tsunami warning including effective combination of different methods for more reliable forecasting.
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.
Realtime observations of vertical/horizontal seafloor movements and sea surface height associated with a huge earthquake are crucial for immediate recognition of its causal fault rupture, so that tsunami early warning can be issued and also the risk of subsequent ruptures can be evaluated. For this purpose, we developed an offshore monitoring system using a moored buoy platform to measure, in realtime, the three observables mentioned above and operated it on a trial basis for a year. While operating the system, GPS-acoustic observation of horizontal movement on the buoy was especially a new challenge. To achieve realtime GPS-acoustic observation under conditions of the limited power supply and narrow bandwidth in satellite communication, we developed special hardware suitable for use on a buoy and software to minimize onboard computational procedures and data transmission. The system functioned properly through the year; 53 regular weekly measurements and 55 on-demand measurements at arbitrary timings. Each measurement consisted of 11 successive acoustic rangings. The buoy tended to drift far from the preferred position for GPS-acoustic measurement, i.e., the center of the seafloor transponder array, due to strong current. The accuracy of the GPS-acoustic positioning achieved ∼46 cm (2σ) even only with “a single ranging” when the buoy was inside the array, while it degraded to ∼1.0 m when the buoy was outside the array. Although the 1.0 m accuracy is a detectable level of possible displacement due to a M8-class earthquake in the source region, further improvement to keep the drifting range smaller despite the current will enhance the utilization of the system.
We examine a method to calculate changes in Coulomb failure stress (ΔCFS) from observed GNSS displacements. The method assumes no stress changes on a horizontal plane and a linear elastic relation between strain and stress, represented by Hooke’s law. The ΔCFS distributions calculated using this method are applied to the 2003 Tokachi-oki and the 2016 Kumamoto earthquakes and they are compared with those using a standard dislocation model buried in an elastic half-space. The results suggest that the ΔCFS distribution at a depth of 10 km in a region far from a deformation source can give a first-order approximation using observations of surface displacements. However, ΔCFS distributions near the source cannot be reproduced by the examined method and need to be evaluated using the standard method. We apply the examined method and GNSS displacement data to calculate ΔCFS on major active faults as well as source faults of large inland earthquakes in southwest Japan for the period 1996-2017. ΔCFS from five large earthquakes, including the 2016 Kumamoto earthquake are separately calculated using the standard method with published fault models. Calculated ΔCFS increases by an order of 10 KPa at most faults over the past 21 years. ΔCFS on the source faults for the 2000 Western Tottori, the 2016 Kumamoto, and the 2016 Central Tottori earthquakes reached a maximum just before their rupture. Coseismic and postseismic deformation of the 2011 Tohoku-oki earthquake accelerated an increase of ΔCFS at some faults, including the source fault of the 2016 Central Tottori earthquake and the Arima-Takatsuki fault zone. The examined method can provide information on the activity of inland earthquakes using contemporarily observed deformation, and can hopefully improve the preparedness for earthquakes.
The 2011 off the Pacific coast of Tohoku earthquake (M9.0) produced up to 1.2 m subsidence along the Pacific coast in northeastern Japan. Based on Global Positioning System (GPS) observations, continuing postseismic coastal uplift has been detected in the past six years after the main shock. By applying a 3D spherical Earth viscoelastic finite element model using the postseismic seafloor and terrestrial GPS observations as constraints, I demonstrate that this uplift is mainly caused by the postseismic viscoelastic relaxation of the asthenosphere. Although the model was constrained only based on horizontal crustal deformation, the vertical displacements predicted for six years after the 2011 Tohoku earthquake agree reasonably well with the time series of the observed uplift at sites along the Pacific coast including the southern Sanriku coast and Kanto district. I estimated the time at which the cumulative postseismic uplift will fully compensate the coseismic subsidence. The results show that large coseismic coastal subsidence on the southern Sanriku coast will be fully offset by the postseismic uplift within several decades. To the immediate north, the model underpredicts the postseismic uplift and possibly indicates unaccounted postseismic fault creep. Farther south, the postseismic uplift of the coast of the Kanto district has already exceeded the small amount of coseismic subsidence over the past six years, as predicted by the model. To prepare for future earthquakes, especially with respect to the coastal construction at fishery ports, it is important to construct a comprehensive rheological structure model based on geophysical observations including GNSS positioning.
Postseismic deformation following the 2011 Tohoku-Oki earthquake has been observed by the Global Navigation Satellite System (GNSS) Earth Observation Network System (GEONET) and the Seafloor Geodetic Observation (SGO) over the past six and half years. Observed deformation at onshore sites exceeds 140 cm horizontally, there is uplift of 50 cm, and deformation tends eastward. However, offshore sites reveal complex patterns ranging from near-zero deformation in the northern part of Iwate-Oki, to westward in the southern part of Iwate-Oki, Miyagi-Oki, and the northern part of Fukushima-Oki regions, and eastward in the southern part of Fukushima-Oki and Ibaraki-Oki regions, respectively. The vertical deformation pattern is more complex than the horizontal. Offshore sites demonstrate subsidence but a large uplift is observed onshore along the Pacific coast. Subsidence is only observed along the Pacific coast in the northern part of Iwate, where there are variations in uplift or subsidence patterns. Many previous 2011 Tohoku-Oki event studies have used a primary model that considers only the afterslip effect. However, westward displacements observed by the SGO highlight the importance of viscoelastic relaxation, even during short-term deformation. It is thus considered that studies on postseismic deformation following the 2011 Tohoku-Oki earthquake should adopt a combined afterslip and viscoelastic model. Postseismic deformation following this event is estimated to continue for more than a few decades; therefore, assessing this effect is crucial for interpreting crustal deformation in Japan. Information on the status of interplate coupling or slip is also vital when assessing earthquake occurrence probability. The continued observation of postseismic deformation and careful monitoring of temporal and spatial changes in interplate coupling or slip will mitigate hazards from successive large megathrust earthquakes and improve understanding of crustal activity in Japan.
In Japan, interplate megathrust earthquakes frequently occur in subduction zones where oceanic plates subduct beneath continental plates, and it is important to elucidate the physical mechanism involved in such earthquakes to prevent associated disasters. Crustal movement data provide essential information to understand plate motion and earthquake source processes. We developed a system that combines GNSS measurements and acoustic ranging techniques to detect seafloor crustal movement. This paper reports the acquisition of recent seafloor crustal movements obtained during campaign observations with a survey vessel, from 2013 to 2016.
A dense Global Navigation Satellite System (GNSS) network has been deployed at Sakurajima volcano since 1995 and extends to the surrounding area of the Aira caldera. The ground deformation obtained by GNSS observation corresponds to transient eruptive activity of Sakurajima volcano, which has produced frequent vulcanian eruptions since 1955. Inflation of the volcano was detected prior to the increase in vulcanian eruptions in 1999, and resumption of the eruptions at the Showa crater. Magma intrusion events and an increase in eruptions in late 2009, late 2011, and early 2015 suggest the existence of an open-conduit system from the Aira caldera to the vents at the summit area of the Minamidake cone, through the sub reservoir beneath the older Kitadake cone. Ground deformation induced by sudden dike intrusion is different from that of previous intrusions, as revealed by the dense GNSS network. GNSS data are useful in evaluating and forecasting volcanic activity, and are available to grasp the advection and diffusion of volcanic ash.
In recent years, earthquake swarm activities have occurred at the Hakone Volcano in the western area of Kanagawa Prefecture, Japan, with a frequency of once in several years. Global Navigation Satellite System (GNSS) observations have detected the inflation of volcanic edifice during these activities. Hot Springs Research Institute of Kanagawa Prefecture (HSRI) regularly observes crustal deformation for monitoring seismic and volcanic activities by using 16 sites of GNSS observation, which were installed in the western area of Kanagawa Prefecture. These observed data, together with those from other agencies, are analyzed routinely, and time-series graphs, displacement vector diagrams, and strain maps are illustrated to monitor seismic and volcanic activities. Given that GNSS monitoring detected the baseline extension about half a month or a month before the earthquake swarm activities, a stacking analysis is routinely performed for early detection of the extension. Some of the analysis results can be found on the website of HSRI. The Hakone Volcano had the largest earthquake swarm activity beginning at the end of April 2015, and a phreatic eruption occurred in Owakudani at the end of June 2015. The GNSS observed crustal deformation, which indicated the inflation of the volcanic edifice in early April 2015. This inflation can be explained by a volume change of a point pressure source located about 6.5 km below sea level.
Two-dimensional ionospheric total electron content (TEC) maps have been derived from ground-based Global Navigation Satellite System (GNSS) receiver networks and applied to studies of various ionospheric disturbances since the mid-1990s. For the purpose of monitoring and researching ionospheric conditions and ionospheric space weather phenomena, we have developed TEC maps of areas over Japan using the dense GNSS network, GNSS Earth Observation NETwork (GEONET), which consists of about 1300 stations and is operated by the Geospatial Information Authority of Japan (GSI). Currently, we are providing high-resolution, two-dimensional maps of absolute TEC, detrended TEC, rate of TEC change index (ROTI), and loss-of-lock on GPS signal over Japan on a real-time basis. Such high-resolution TEC maps using dense GNSS receiver networks are one of the most effective ways to observe, on a scale of several 100 km to 1000 km, ionospheric variations caused by traveling ionospheric disturbances and/or equatorial plasma bubbles, which can degrade single-frequency and differential GNSS positioning/navigation. We have collected all the available GNSS receiver data in the world to expand the TEC observation area. Currently, however, dense GNSS receiver networks are available in only limited areas, such as Japan, North America, and Europe. To expand the two-dimensional TEC observation with high resolution, we have conducted the Dense Regional and Worldwide International GNSS TEC observation (DRAWING-TEC) project, which is engaged in three activities: (1) standardizing GNSS-TEC data, (2) developing a new high-resolution TEC mapping technique, and (3) sharing the standardized TEC data or the information of GNSS receiver network. We have developed a new standardized TEC format, GNSS-TEC EXchange (GTEX), which is included in the Formatted Tables of ITU-R SG 3 Databanks related to Recommendation ITU-R P.311. Sharing the GTEX TEC data would be easier than sharing the GPS/GNSS data among those in the international ionospheric researcher community. The DRAWING-TEC project would promote studies of medium-scale ionospheric variations and their effect on GNSS.
In recent years, the role of tourism-related workers in regional volcanic disaster prevention has increased in Japan. The coexistence of tourism with disaster mitigation is important in keeping residents and visitors safe and in protecting livelihoods. This paper analyzes responses from tourists and tourism workers on their receipt of volcanic hazard information. Awareness of this hazard is developing in the tourism industry. Information of expert such as members of the JMA and volcanologists at universities and institutes were considered more reliable sources of information than others. However, a direct access to experts’ information was not considered easy. Respondents’ recognition of the past hazards of Zao Volcano and future hazard factors were almost accurate. Some tourism-related workers hoped to obtain volcanic hazard information from the experts to provide to their customers. Many respondents had excessive expectations for predicting an eruption. A few were able to accept the uncertainties associated with volcano warnings and status reports. Experts need to provide adequate explanations of scientific evidence and the associated scientific uncertainties before society can readily accept eruption warnings. Furthermore, in an emergency, it is necessary to make available accurate information from specialized agencies and experts, and promptly provide them to tourism companies.
The effects of framing are important to the perception of earthquake risk. This research investigated the effects of framing on the people’s perception of earthquake risk in Chiang Rai, Thailand. There were three frames conveying the same earthquake risk but presented in different terms. These statements were in frequency terms of building damage: 475 severely damaged private buildings in 500 years, 10% chance of occurrence in 50 years, and one severely damaged private building per year. The objective of this research was to determine whether framing the same earthquake risk in different terms led to different perceptions of the risk by different people, leading to 1. what would be the most effective framing type regarding severely damaged private buildings, 2. which affected people’s earthquake risk perception most, and 3. whether experiencing an earthquake disaster would change the risk perception of the residents. The result showed that presenting the risk as “one severely damaged private building per year” was the most commonly selected among the three frames and was statistically significant. This finding clarifies that short time frames influence people’s earthquake risk decisions, and agencies can effectively use framing to communicate the earthquake risk to people in seismic regions to stimulate their earthquake preparedness and to reduce future earthquake risk. However, earthquake experience does not change the risk perception of the residents.
In the field of disaster prevention, disaster loss is often classified into “direct loss” and “indirect loss.” As such, “total loss” is often calculated as a sum of “direct loss” and “indirect loss,” where “direct loss” is defined as a “loss of capital (assets) as a stock” and “indirect loss” is defined as “loss arising out of decline in postdisaster production as a flow.” However, the loss here is calculated twice. The calculation is incorrect if “indirect loss” refers to, in particular, the lost profit of a firm that has lost a production facility that is considered as a stock. The reason is that the “value of capital stock” is nothing but the present value of a product that the stock will produce in the future. Therefore, an “indirect loss” defined in the above manner corresponds to a decrease in stock value. Using a dynamic economic model, this article provides a basic structure, “value of loss in capital stock lost by a disaster” = “total decline of production after a disaster.” This article also presents a relational expression in consideration of the restoration cost of a production facility, and concludes that a more multifaceted and functional damage information system needs to be developed in the future.
Mr. Yokomatsu provided valuable comments on the paper JDR Vol.11 No.6, pp. 1190-1201 by Takeuchi and Tanaka. The discusser claims that if “direct loss” is defined as a “loss of assets (capital stock)” and “indirect loss” as a “loss of postdisaster production decline,” then the sum of the two as the “total loss” of a disaster contains a double count of disaster loss, because the “value of assets (capital stock)” for “direct loss” is the present value of an output to be produced in the future by the assets, which is nothing but “indirect loss.” This is a misunderstanding caused by an ambiguous definition of the term “loss of assets.” In order to avoid such misunderstanding, the authors clarify the definition by elaborating the content of “total loss” = “direct loss” + “indirect loss” as the “total loss of assets” = “loss of the physical acquisition cost of assets” + “loss of the potential production profit of assets.”