In March 2015, the Third World Conference on Disaster Risk Reduction adopted the Sendai Framework for Disaster Risk Reduction with a two-part goal: to prevent new and reduce existing disaster risks through the implementation of integrated and inclusive measures that prevent and reduce hazard exposure and vulnerability to disaster, and to increase preparedness for response and recovery, thus strengthening resilience. The first priority for action was given to ”understanding disaster risk,” including focusing on the collection and use of data, risk assessment, disaster prevention education, and awareness raising. The stance of emphasizing science and technology was clearly expressed.

In September 2015, the UN Summit meeting adopted the 17 goals of the 2030 Agenda for Sustainable Development. Four of the 17 goals include targets related to disaster prevention and mitigation, which has given rise to active discussions over measurement methods and indicators for the targets. The Paris Conference of the UN Climate Change Conference (COP21), held from the end of November to early December 2015, placed an emphasis on the importance of science and technology in both mitigation and adaptation.

In light of these international discussions and their outcomes, we called for papers on the following three topics for this special edition featuring water disasters.

(1) Prevention of new water disaster risks: rainfall prediction, flood and drought prediction, river bed change prediction, climate change, land use plans, etc.

(2) Reduction of existing water disaster risks: disaster data and statistics, risk monitoring, risk assessment, etc.

(3) Resilience reinforcement and inclusive measures: disaster recovery, risk communication, competence development, etc.

Nineteen papers were applied to this special issue. All papers were peer reviewed, and sixteen papers are included herein. We received invaluable comments and suggestions for all applications from the points of view of various fields from many experts in Japan and overseas. We would like to express our gratitude for these.

In order to be able to issue flood warnings not hours but days in advance, numerical weather prediction (NWP) is essential to the forecasting of flood-producing rainfall. The regional ensemble prediction system (EPS), advanced NWP on a local scale, has a high potential to improve flood forecasting through the quantitative prediction of precipitation. In this study, the predictability of floods using the ensemble flood forecasting system, which is composed of regional EPS and a distributed hydrological model, was investigated. Two flood events which took place in a small basin in Japan in 2010 and which were caused by typhoons Talas and Roke were examined. As the forecasting system predicted the probability of flood occurrence at least 24 h beforehand in the case of both typhoons, these forecasts were better than deterministic forecasts. However, the system underestimated the peak of the flooding in the typhoon Roke event, and it was too early in its prediction of the appearance of the peak of the flooding in the Talas event. Although the system has its limitations, it has proved to have the potential to produce early flood warnings.

In investigating glacier mass balance and water balance at Huayna Potosi West, a glacierized basin in the Bolivian Andes (Cordillera Real), we used a remote sensing method with empirical area-volume relationships, a hydrological method with runoff coefficients, and water balance method. Results suggest that remote sensing method based on the glacier area from satellite images and area-volume relationships is too imprecise to use in performing analysis in short time intervals. Glacier mass balance obtained using a new area-volume relationship was, however, similar to that obtained by the water balance method, thus proving that the new area-volume relationship is reasonable to use for analyzing glaciers within a certain size range. The hydrological method with a runoff coefficient considered glacier as the only storage for saving or contributing to runoff and nonglacier area as the only source of evaporation. We applied a fixed runoff coefficient of 0.8 without considering wet or dry seasons in nonglacier areas – a method thus sensitive to meteorological and hydrological data. We also did not consider glacier sublimation. The water balance method is applicable to the study region and excelled other methods in terms of resolution, having no empirical coefficients, and considering sublimation and evaporation. Among its few limitations are possibly underestimating evaporation and runoff over nonglacier areas during wet months and thus possibly overestimating glacier contribution at mean time.

Despite recent advances in hydrological models and observation technology, the prediction of floods using advanced models and data has not yet been fully implemented for practical use. The major issues in prediction originate from the underlying uncertainty of the initial conditions of the basin and the accuracy of the precipitation forecast. Effective transmission of flood information to corresponding authorities is also necessary when considering countermeasures against an oncoming flood. We present in this article a data archive and model integrated system to overcome these issues. The system realizes flood forecasting by employing a land surface model coupled with hydrological model and an ensemble precipitation forecast model to address the accuracy of initial conditions and precipitation. While the Water and Energy Budget Based Distributed Hydrological Model (WEB-DHM) rigorously estimates the physical state of the basin, the ensemble precipitation forecast model analyzes historical errors in forecasts and returns precipitation ensembles reflecting the uncertainty in the forecast specifically regarding the target basin. A combination of these models yields an ensemble of streamflow forecasts. We further develop a virtual reservoir simulator to enhance the proactive use of forecast information to support decision-making by reservoir managers. These models are integrated into the Data Integration Analysis System (DIAS). The feasibility of the system for practical use is tested against data from recent typhoon events.

This study simulated the inundation process in the Lower Mekong River Basin (LMB). The LMB has suffered from severe floods, especially in 2000 and 2011. To quantify the inundation of water in a basin where large-scale inundation by river water occurs, understanding the conveyance of a river channel during a flood is particularly important. Therefore, we conducted a field survey using an acoustic Doppler current profiler (aDcp) to understand the longitudinal distribution of the width and depth of the river channel and the variation in hydraulic resistance with respect to shear stress on the riverbed. It was found that the width and depth vary longitudinally, and the relationship between them can be estimated by an equation derived from governing equations of water and sediment and the bed load formula. Furthermore, it was revealed that hydraulic resistance decreases with increasing non-dimensional shear stress. Then, the characteristics of the river channel were incorporated into the runoff-inundation simulation. Furthermore, inundation water should be validated not only in terms of inundation extent but also with respect to water depth and velocity. These were estimated using 8-day composite surface reflectance data from the Moderate Resolution Imaging Spectrometer (MODIS) and the SRTM. Simulation results indicated that water level and discharge within the river channel were able to reproduce observed values. Additionally, simulated inundation extent, water velocity, and water depth over the floodplain showed reasonable agreement with the results using the data from the MODIS and the SRTM. Although there are some elements that should be improved, the inundation process in the LMB was simulated appropriately despite its complexity. The method described in this study to set a calculation condition and to validate variables over a floodplain should be useful for runoff-inundation simulation in various large-scale basins.

A method is proposed to predict bank erosion and sand bar migration in river reaches where suspended sediment transport is dominant. The method focuses on the influence of the lateral bed slope on the erosion and deposition rate of suspended sediment, as well as on the profile of lateral bedload transport, assuming that geometric similarity holds in the bank region. In the proposed model, the erosion and deposition rate can be evaluated using either the bed shear stress at a reference location or the average bed shear stress in the bank region. In order to simulate bank erosion and associated bank shifting with a depth-integrated-base treatment, stretchable grids were added to the conventional coarse grid system near the bank. The proposed method, including the bank erosion model, is applied to the lower reach of the Brahmaputra River, which is ∼90 km long and ∼12.50 km wide. The computed results on bank shifting, sand bar migration, and sediment transport rates are compared with data obtained from field investigations and remote sensing. These results suggest that the proposed method is applicable for predicting sediment issues in river reaches dominated by suspended sediment.

The standardized precipitation index (SPI) has been used to monitor and analyze meteorological droughts using long-term monthly precipitation from national meteorological and hydrological services on multiple timescales. Instead of evaluating climatic impacts with separately-computed SPI for present and future climates, we introduced the comparative SPI (cSPI) computed using target (future) datasets on the basis of a reference (present) dataset. The cSPI approach evaluates standardized precipitation change in one dataset for different periods and for different datasets in a common period. Using 12-month cSPI, we investigate the change in central conditions and in the probabilities of dry and wet conditions between present and future climates. Meteorological drought and flood hazards in Asia are examined with MRI-AGCM3.2S, a 20-km mesh global atmospheric model, time-slice experiments of the present (1979–2003) and future (2075–2099) with four different sea surface temperature patterns. As one result indicates, the median of the 12-month cSPI shifts to severely dry around the Mediterranean Sea to the Persian Gulf, and to extremely wet in the Tibetan Plateau, North and South India, and around the Yellow Sea. Therefore, we conclude that the cSPI approach is a useful way to characterize both future drought and flood hazards under climate change.

Drought is a slow-developing disaster of water shortages in water cycle components adversely affecting anthropogenic water use. This study introduces a drought assessment framework of standardized indices in Pampanga (Philippines), Solo (Indonesia), and Chao Phraya (Thailand) basins. We used three existing and developed two new standardized indices to characterize meteorological, agricultural, hydrological and socio-economic droughts. We constructed a 15-arcsec (about 0.45-km) grid block-wise TOP (BTOP) model with multipurpose dam operation at individual river basin using global datasets and calibrated BTOP models with daily river discharge and dam inflow data. The simulated irrigated area is also compared with historical drought damages at each river basin. The calibrated BTOP models were run with bias-corrected MRI-AGCM3.2S precipitation to evaluate droughts under climate change. The calculated standardized indices show similar drought timing of the 1982-1983, 1987-1988, 1991-1992, 1997-1998 and 2002-2003 droughts across three river basins. In addition, the timing of these droughts coincides with historical El Niño-Southern Oscillation (ENSO) cycle events. The projected future climates demonstrate a variability of dam inflows and drought severities between four cases of the worst (RCP8.5) climate change scenario. We conclude that standardized indices are useful tools to characterize droughts at water cycle components.

The present study demonstrates a method to specify critical rainfall conditions for the occurrence of a sediment disaster and identify areas prone to landslides using a simulator proposed by the current authors for sediment hazards. The simulator predicts spatial and temporal distributions for surface and subsurface flows, landslides, and debris flow resulting from rainfall events. The method to develop a critical curve for the occurrence of a disaster is proposed using simulated landslide data derived from artificially specified rainfall conditions, past rainfall data, and disaster records. Usually, a rainfall event also constitutes a period without rain, and this method can be used to evaluate the influence of the no-rain period. In addition, we propose a method to classify slopes according to the probability of landslide occurrences on a domain defined by slope gradient versus catchment area, using data on landslides resulting from a specified rainfall amount and intensity. Areas identified as having a high probability of landslide occurrences correspond to the runout mark of landslides and debris flow in August 2014.

Heavy rainfall in September 2015 inundated the Kinu River basin and caused an overflow and dike breach of the river in the eastern part of Joso City, Ibaraki Prefecture, Japan. The area, which is bounded by the Kinu and Kokai Rivers, experienced deep inundation of more than 2.5 m, which continued for more than 3 days at the maximum level. Although the estimated maximum inundation depth and duration were basically related to the elevation and distribution of topographic surfaces, strong flood flow washed buildings away near the overflow and dike breach sites and caused deep inundation even though these sites were located on a relatively higher natural levee. In addition, serious damages such as interruption of emergency transportation routes and deep inundation over floor level occurred, isolating evacuation centers and important facilities including a municipal hall and hospitals. Few residents utilized the pre-prepared flood hazard map or understood the local geography, and evacuation orders were not fully transmitted to the local residents, which might have increased the flood’s impact.

Globally, large-scale floods are one of the most serious disasters, considering increased frequency and intensity of heavy rainfall. This is not only a domestic problem but also an international water issue related to transboundary rivers in terms of global river flood risk assessment. The purpose of this study is to propose a rapid flood hazard model as a methodological possibility to be used on a global scale, which uses flood inundation depth and works reasonably despite low data availability. The method is designed to effectively simplify complexities involving hydrological and topographical variables in a flood risk-prone area when applied in an integrated global flood risk assessment framework. The model was used to evaluate flood hazard and exposure through pixel-based comparison in the case of extreme flood events caused by an annual maximum daily river discharge of 1/50 probability of occurrence under the condition of climate change between two periods, Present (daily data from 1980 to 2004) and Future (daily data from 2075 to 2099). As preliminary results, the maximum potential extent of inundation area and the maximum number of affected people show an upward trend in Present and Future.

Flood damage to agriculture (rice crops) was assessed in the Pampanga River basin of the Philippines. Flood damage to agriculture was defined as a function of hazard characteristics, such as flood depth and flood duration, exposure, and growth stage of rice crops, and estimated in terms of yield loss using a depth-duration-damage function. The assessment of flood damage to agriculture in the Pampanga River basin was conducted using Digital Elevation Model data of the Interferometric Synthetic Aperture Radar (IfSAR-DEM) and Digital Elevation Model data of the Shuttle Radar Topography Mission (SRTM-DEM). The results were further improved using highly accurate IfSAR-DEM. To assess flood disaster damage, a hazard assessment was conducted using the Rainfall Runoff Inundation model. Estimated values from the agricultural damage assessment during the flood event from September 26 to October 4, 2011 were compared with reported values. The accuracy of flood hazard assessment and flood disaster risk assessment highly depends on the quality of topographical data, and better results can be obtained by using highly precise topographical data. Flood disaster risk assessment in the agricultural sector was also conducted for a recent flood in October 2015 and flood events with different return periods of 10, 25, 50, and 100 years. The assessment results based on the different return periods of flood events were then used to estimate the probability of agricultural damage for most frequently damaged and rarely damaged areas. The results of flood damage assessment in the Pampanga River basin provide a basis to identify areas at risk, and these results can be useful for planners, developers, policy makers, and decision makers in establishing policies required to reduce flood damage.

Disaster risk assessment is vital to determining needs for disaster countermeasures and promoting their implementation. However, it is difficult to conduct evidence-based risk assessment in flood-prone areas of Asia due to area-specific characteristics such as limited local data on natural and societal conditions and local lifestyles of persons who have adapted to frequent floods. This paper proposes basic flood risk assessment considering these characteristics and explores a case study conducted in a flood-prone area of the Pampanga River basin in the Republic of the Philippines to verify our method. We surveyed local household members as part of the study to understand local situations, finding that past flood damage cost little thanks to building structures adapted to frequent flooding and to local ways of protecting property during floods. We also found that the use of depth-damage curves developed for urban areas may overestimate anticipated damage expected in future floods when these curves are applied to flood-prone rural areas. For this reason, we propose a method of flood risk assessment for evaluating the societal impact on residents’ lives using observed thresholds of inundation depth by flood simulation, rather than using a method that estimates damage cost. Application of our proposal to the case study area confirmed its applicability and effectiveness in evidence-based planning for reducing flood risk.

This paper proposes a method to evaluate the flood risk of each district in a municipality to assist disaster management personnel. The method is specifically for municipalities in a mountainous region where insufficient information is available for practical disaster management. Using this method, we conducted inundation analysis for multiple patterns of rainfall and discharge using a rainfall-runoff-inundation model, and estimated the maximum inundation depth and duration. Based on the estimation, we developed a “flood diagnostic chart” to evaluate district-level flood risk, additionally considering other indicators. Moreover, we located flood hotspots, which are areas requiring extra precautions because of the high flood risk for districts.

The flood in the Chao Phraya River basin in 2011 caused 815 deaths and more than $45 billion in economic damage to Thailand. The industrial sector, the main contributor to the country’s economy, suffered especially devastating economic damage due to inundation of the industrial areas in the river basin. Seven industrial areas, where the total share of Japanese-affiliated firms was more than half, were most severely hit by the flood. In this study, a survey was conducted with Japanese-affiliated firms from February–March 2015 with the purpose of further strengthening their flood countermeasures in the future. The firms’ factories were asked which lessons from the 2011 flood they considered important and whether their experiences from the flood had been applied to strengthening their flood countermeasures. It was found that these factories, regardless of their inundation status in 2011, considered “preparation of a business continuity plan or manual for possible floods” to be the most important lesson from the 2011 flood. However, when it comes to actual implementation of flood countermeasures, the factories that were inundated in 2011 strengthened flood countermeasures much more than the factories with only indirect damage. In both groups, however, collaboration with business partners and local communities as well as use of reliable flood-related information were revealed to be the areas where further strengthening would be possible.

Banda Aceh, Indonesia, Tohoku cities in Japan and Tacloban, Philippines were all completely destroyed and have recovered or are now recovering from the ocean of debris. Banda Aceh and Tacloban have recovered to a normal state rather quickly within two years or so after the disaster’s occurrence. The Tohoku cities are taking a much longer time and even now, more than 170,000 (March 10, 2016) people are in evacuation houses of various kinds. Such a difference comes from the basic selection of the recovery process, based on the basic policy of reconstruction.

Building resilient cities is one of the Sustainable Development Goals with disaster risk reduction targets. In order to build resilient cities, the strategy of building back better, a new focus priority in the Sendai Framework for Disaster Risk Reduction, plays a key role. As disasters occur everywhere one after another, recovery processes also take place everywhere after each disaster. Building back better is therefore one of the most practical ways of building resilient cities.

Quick recovery has many advantages if it extends to building back better toward resilient cities but in almost all cases as experienced in Banda Aceh and Tacloban, once a city is recovered freely, it is extremely difficult to redesign and gradually install resilience into the city formation. On the other hand, slow recovery and waiting time, as experiencing in Tohoku cities, make people suffer, make local economies difficult to recover, and have high national costs. It is difficult to assess how and under what conditions the cost of such investment may be recovered by building resilient cities with long-term safety.

The justification for selecting a recovery trajectory depends on the state of the national economy as well as the safety culture of the nation. Yet more important and practical support for building back better is having a pre-disaster recovery plan prepared before a disaster occurs. In fact, regardless of the availability of official pre-disaster plans, the redevelopment and reform efforts to improve communities in normal times will help promote a swift and effective reconstruction when an unexpected disaster occurs. This was experienced in Tokyo after the Great Kanto Earthquake and after World War II, as well as in many cities in Japan.

Practical flood management depends on the extent to which the cost of taking risks is shared in a society among governments, interested parties, communities, and individuals. Risk management calls for identification, analysis, assessment, control, avoidance, minimization, or elimination of unacceptable risks through policies, procedures, and practices under three strategies for risk management: reduction, retention, and transfer. Flood risk management under a variety of uncertainties, such as the impacts of climate change, favors the implementation of flexible and adaptive management in top-down and bottom-up approaches. The former uses projections of global or spatially downscaled models to drive resource models and project impacts. The latter uses policy or planning tools to identify which changes in climate would most threaten their long-range plans or operations. Particularly for the bottom-up approaches, appropriate indicators that directly assess the flood risk of each area are essential. This study analyzes the international efforts of robust flood management from the top-down and bottom-up approaches, such as insurance and indicator and management systems, to seek incentive mechanisms for risk management. To implement international commitments, such as the Sendai Framework for Disaster Risk Reduction, there are gaps in implementing a holistic approach to flood management strategies and, therefore, mainstreaming disaster risk reduction and addressing sustainable development. The robustness of flood management requires the capacity-building necessary to understand and sufficiently respond to flood hazards, vulnerabilities, and benefits as an important evolutionary link in the transition between implementing global development goals, such as the Sustainable Development Goals, and disaster risk-reduction activities. There are three challenges: data and information infrastructure, inter-disciplinary knowledge development, and trans-disciplinary policy.

The Standing Technical Committees on Disaster Risk Management (CDRM) of the World Federation of Engineering Organizations (WFEO) play an important role in collecting and disseminating DRM-related information and knowledge that will conceivably help engineering society members take effective disaster mitigation measures. As part of achieving this mission, the CDRM conducted two important 2015 events – the WFEO-CDRM Special Session on Disaster Risk Management at the 11^th International Conference of the International Institute for Infrastructure Resilience and Reconstruction (I3R2) (I3R2 session) held in Seoul, Korea, and the 9^th Joint International Symposium on Disaster Risk Management conducted in conjunction with the International Symposium on River Technologies for Innovations and Social Systems held in the 2015 World Engineering Conference and Convention (WECC2015) in Kyoto, Japan (WECC2015 symposium).

The I3R2 session featured seven presentations. During the first half, disaster-cause papers covered high typhoon tides, earthquakes, and rain-induced soil erosion. The second half focused on mitigation-measure presentations such as recovery/reconstruction and regional support for mothers and children in the event of disasters.

The WECC2015 symposium featured ten presentations by ten speakers with widely varied backgrounds in disaster mitigation, river engineering, international cooperation, UNESCO regional centers, NPO management, science and technology sections at embassies, and ferry and resort complex management. These informative, meaningful presentations close with active and informative Q&A sessions.

In this special issue, five presentations that were revised as a form of academic paper were selected and published. I hope that these papers will be utilized for further advancement of disaster mitigation measures.

The Great East Japan Earthquake Tsunami on March 11, 2011, caused unprecedented damage mainly in northeast Japan. This paper introduces the characteristics of the tsunami and resultant damage. The mechanism of the breaching of coastal structures and the effect of surviving structures on damage reduction in land are discussed. Then, a two-level tsunami mitigation concept, proposed and adopted by a committee hosted by the Government for recovery and reconstruction, is introduced. Within this framework, coastal structures are required to be resilient to external forces exceeding the design level. Various technologies that have been developed based on experience and research after the tsunami are introduced.

The Ariake Sea has Japan’s largest tidal range – up to six meters. Given previous Ariake Sea disasters caused by storm surges and high waves, it is considered highly likely that the bay’s innermost coast will be damaged by typhoon-triggered storm surges. Concern with increased storm-surge-related disasters is associated with rising sea levels and increasing typhoon intensity due to global warming. As increasingly more potentially disastrous typhoons cross the area, preventing coastal disasters has become increasingly important. The first step toward doing so is damage prediction, which requires numerical simulation. Our study considers the tracks of typhoons considerably influencing the Ariake Sea. To examine storm-surge risk related to both inundation area and process, we calculated storm surges inundating the Sea’s innermost coastal area using an improved ocean-flow finite-volume coastal ocean model. Results showed that enhanced storm surges were to be anticipated and that inundation areas could be extensive where typhoons followed a route from west to northeast across the Sea. We also found that even under current climatic conditions, typhoons able to cause significant storm-surge and inundation disasters could adversely affect the Bay’s innermost coastal area. Our analysis of this area and process indicated that the inundation extent around the bay’s innermost coast varies with the typhoon, confirming the importance of determining typhoon routes triggering the potentially greatest inundation damage.

Soil-organic amendment (SOA) is one of the sustainable soil improvement measures to mitigate climate change related issues such as rainfall-induced hazard and soil erosion. Organic wastes particularly compost and biochar can be reused and recycled into viable resources. However, there are limited data on incoporating organic wastes into a soil that is susceptible to erosion by rainfall. Therefore, objective of this study is to investigate properties of a soil from Okinawa prefecture (Kunigami maaji) that are associated with resisting ability against artificial rainfall intensities of 30, 60, 90 and 120 mm/h after adding two organic matters: household-derived compost and rice hush-derived biochar. The properties were soil-water retention, runoff, soil loss, infiltration and electrical conductivity. The compost was mixed with the soil at application rates of 0.5, 1.0, 1.5 and 2.0 kg/m². The compost of 1.0 kg/m² was mixed with the soil and the biochar at application rates of 1, 3, and 5% by total weight. Experimental results indicate that the soil water retention properties of the soil were improved by the treatment of compost and biochar. However, soil loss was not significantly reduced under initially saturated soil condition, applied rainfall intensities, testing duration and experimental conditions. The results of this study could be used as baseline data for evaluating correlation between properties of soil water retention curves to soil erosion.

Losses and damages caused by natural disasters have negatively impacted poverty alleviation and human development and undermine the achievement of the Millennium Development Goals (MDGs). However, disaster issues were not included in MDG targets set up in 2000. A new development agenda, Sustainable Development Goals (SDGs), was approved in the UN General Assembly in September 2015. In the SDGs, disaster issues are included in many targets such as target 11.5. To appropriately set targets and prepare monitoring measures for disaster-related issues, quantitatively measurable indicators of impacts of disaster risk reduction on economic growth and poverty alleviation should be prepared. In addition, to promote disaster prevention measures, we need to convince policy makers that such measures are highly essential for a country’s development and are cost-effective. Although the cost-effectiveness of single disaster prevention projects has been studied, aggregate effectiveness of these projects at a national level has not been presented. This study proposes a simple method to explain the cost-effectiveness of flood protection investment in Japan post World War II by using national aggregate data.

Numerous landslides due to torrential rain have frequently occurred in the urban fringe of many cities in recent years. To protect residents from a landslide, there is a need that those living in landslide danger zones migrate to safer areas. In this study, we examined the financial feasibility of a scheme that enables residents in landslide danger zones to migrate to safe residential areas. The case study of the Kyushu region indicated that relocation for disaster adaptation is financially feasible for 20% of landslide danger zones in the region.

In this study, empirical fragility curves expressed in terms of relationship between damage ratio indices of buildings and ground motion indices were developed in northern Miyagi prefecture located in near-field areas during the 2011 off the Pacific coast of Tohoku Earthquake. The ground motion indices were evaluated from observed ground motions at strong-motion stations and estimated at sites at which no strong-motion accelerometers were deployed during the mainshock. The ground motions at the non-instrumental sites were estimated using the empirical Green’s function method based on bedrock motions inverted from observed records on surfaces from small events that occurred inside the source fault, transfer functions due to underground velocity structures identified from microtremor H/V spectral ratios, and a short-period source model of the mainshock. The findings indicated that the empirical fragility curves as functions of Japan Meteorological Agency (JMA) instrumental seismic intensity during the 2011 Tohoku Earthquake almost corresponded to those during the 1995 Kobe Earthquake and the seven disastrous earthquakes that occurred between 2003 and 2008. However, the empirical fragility curves as functions of peak ground velocity were the lowest. A possible reason for this is that the response spectra of the ground motions in the period ranging from 1.0 s to 1.5 s were small during the 2011 Tohoku Earthquake. Another reason could be the seismic resistant capacities of buildings in the studied districts involved during the 2011 Tohoku Earthquake exceeded those in the cities affected during the 1995 Kobe Earthquake.

Natural disasters recently occurred in the northern region of Thailand have been increasingly becoming an important issue with emphasis on the alarm and caution for damage and frequency of disasters which may cause major losses of human lives and properties. In May 2014, six provinces in the northern region of Thailand were affected by the earthquakes (the main shock of Mw 6.1 (Moment Magnitude by USGS) and hundreds of aftershocks including the one Mw 5.9 at most) that had the epicenters at Phan District, Chiang Rai Province (19.656°N 99.670°E). This research aimed to study the distribution of seismic accelerations at Mae Ngad Somboon Chon Dam in Chiang Mai. The accelerometers were installed at 3 positions in the dam, consisting of the crest, middle and base of the dam. The collected data were compared and analyzed by the Finite Element Method. Analyses of the Linear Elastic Model, the Mohr-Coulomb Model and the Hardening Soil Model with using the acceleration actually recorded at the dam were conducted to determine an appropriate analytical model. The results indicated that the accelerations obtained from the Hardening Soil Model were more suitable for actual accelerations among others. Therefore the behavior of the Hardening Soil Model is more realistic than that of the Linear Elastic Model or the Mohr-Coulomb Model.