Long-term in-situ pressure gauge data and nearby tide gauge data collected at locations affected by the Kuroshio are examined. The major tidal components can be removed from the in-situ pressure data by calculating the moving average for one-month or longer. An anomalously high sea level was observed from 2004 to 2005 by the tide gauges and is attributable to the warm water intrusion into the Sagami Bay caused by the Kuroshio large meander. Meanwhile, the warm water contributed little to the pressure change. No significant influence of the large meander was identified in the data of the in-situ pressure gauge and the tide gauge at Murotomisaki, where the horizontal displacement of the Kuroshio current path was small.

 A number of coastal forests were heavily damaged by the 2011 Great East Japan Earthquake (GEJE) tsunami. Since the resulting wood debris potentially increased the damage in some coastal communities, it is essential to elucidate the limitation of coastal forests as a tsunami mitigation measure. Although some criteria based on a maximum inundation depth have been proposed, their applicability has not been thoroughly examined. Moreover, some case studies following the event suggested that explanatory varia-bles other than inundation depth, i.e., local topographic features around coastal forests, significantly af-fected the damage level of the coastal forests. This study aims to identify potential variables and statisti-cally relate them to the damage of coastal forests based on a numerical tsunami simulation and a dataset of forest damage collected after GEJE tsunami. Consequently, the followings were confirmed: 1) the probability of severe damage on coastal forests increases with the fluid force of tsunamis and an elevation gap in front of the forest due to seawalls or other local landforms, 2) the probability of severe damage de-creases with the elevation of the coastal forest itself. Additionally, a damage prediction model was con-structed based on the results above, which achieved higher accuracy than the existing method.

 Coastal vegetation, such as coastal forests and mangroves, has been focused on its mitigation effects as green infrastructure. However, the current numerical calculation speeds do not allow for detailed modeling of vegetation with complex root structures. In this study, we aimed to establish a low-cost calculation method for mangroves by using the porosity. The validity of the method was verified by comparing with physical experiments, and the disaster mitigation effect of a raised road was examined using the cross-sectional shape of the local site. The numerical results are consistent with the experimental results, although they are slightly underestimated.

 Artificial reefs have been evaluated as facilities which can provide disaster prevention functions while minimizing environmental impacts, but there have been few cases where artificial reefs have been expected to actively create a favorable environment. The artificial reef at Hamasaki Coast in Motobu Town, Okinawa, has been shown to have much better coral establishment and growth conditions than surrounding artificial structures. However, much remains to be learned about the relationship between growth conditions and physical environment. The purpose of this study is to discuss the relationship between coral distribution characteristics and the physical environment of current, water quality and sediment by conducting field surveys on the artificial reef at Hamasaki Coast, and to identify desirable environmental factors for coral growth. An analysis of the relationship between flow conditions, sediment volume, and coral distribution showed that coral growth is related to the creation of an environment in which suspended sediment is difficult to accumulate due to fast currents during low and high tides.

 To analyze the mangroves distribution change in Aira river mouth in Iriomote Island and the effects of sea level rise, the growing ground heights of four mangroves species were measured. As a result of analyzing the tidal observation records in Ishigaki Island, the sea level rise rate was about 2.3 mm/year, and the annual inundation period of the ground height where the tallest mangroves were growing was 3288 hours for A. marina, 2424 hours for S. alba, 2040 hours for R. stylosa, and 1584 hours for B. gymnorhiza. As a result of analyzing the images of Aira river mouth taken in the past, the mangrove forest expanded from 5.24 ha in 1978 to 7.74 ha in 2017, and the R. stylosa forest transitioned to the B. gymnorhiza forest, and the S. alba growing on the tidal flat to the R. stylosa forest. By analyzing the tidal observation records in Ishigaki island and the estimated distribution of mangrove species in Aira river mouth, the sediment accumulation rate of mangrove species in Aira river mouth is about 3.85 mm/year, which is higher than the sea level rise rate of 2.3 mm/year, indicating that the distribution of mangrove species has been expanding.

 The Yatsu tidal flats, located in the inner part of Tokyo Bay, are among the most significant tidal flats remaining in Japan. The tidal flats were registered as a Ramsar site in 1993 because of the numerous migratory birds that visit the wetland annually to breed and feed. In recent years, the occurrence of green tides due to the extraordinary growth of Ulva spp. has become problematic. Since 1995, the green tides have severely impacted the ecology of the area. However, the green tides disappeared suddenly in 2017, and did not occur again until 2020. Although numerous models have been developed to predict the occurrence of green tides, none of these models have been sufficiently robust. In this study, as a follow-up to a previous study examining the causes of extinction in 2017-2018, the reasons underlying the disappearance of green tides from 2019 to 2020 in the Yatsu tidal flats were examined based on field observations. We also constructed a green-tide prediction model using artificial intelligence and a neural network. The results showed that the disappearance of green tides from 2019 to 2020 was due to insufficient illuminance due to a long rainy season as well as high water temperatures. The predicted values for the green tide area generated by the model closely corroborated actual observations in the field. In addition, water temperature, nitrogen, and phosphorus greatly affected the incidence of green tides at the Yatsu tidal flats.

 This study proposes and projects the new framework for calculating the climatic Maximum Potential Storm surge height (MPS) based on the typhoon’s Maximum Potential Intensity (MPI) theory for Japan’s major 53 bays. The MPS is optimaized based on dynamic storm surge simulation and is estimate the characteristics of wind-surge for each bay. We analyzed future changes in MPS in major 53 bays in Japan as well as MPI in the Western North Pacific (WNP). MPI in the WNP will be intensified by 8.9 hPa for RCP8.5 scenario. Related to the MPI change, MPS in the bays where higher storm surge cases occurred in the past will become higher in the future.

 The compounding effects of storm surges and sea level rise (SLR) are expected to accelerate coastal hazards and inundation. Therefore, projecting the compound effects in coastal territories is important for impact assessments and adaptation. Small island states in the South Pacific are one of the areas that are the most highly impacted by climate change. Storm surges are caused mostly by tropical cyclones (TCs), which may result in a significant portion of the population being temporarily displaced by infrastructure damage, flooding and erosion. Meanwhile, SLR is changing coastlines permanently and may result in a lasting displacement of the exposed population. This study focuses on Viti Levu, the largest and most populous island in Fiji, and estimates the impact of SLR and storm surge on the coasts and population, by using data from observations and future projections to estimate the angle of approach, occurrence frequency, and intensity of TCs. Based on the findings of this study it can be inferred that while Fiji has had a long record of TCs that hit its islands, SLR may become a more significant phenomenon that shapes climate-related migration. While a milder increase in the climate radiative forcing (+2K scenario) will increase the exposed population less dramatically compared with a sharper increase (+4K scenario), we conclude that either scenario would result in a significantly higher number of the population exposed during future storms.

 For coastal adaptation purposes, it is important to estimate climate change related future changes in extreme sea levels due to storm surge and ocean waves, in addition to sea level rise. In this study, global storm surge and ocean wave simulations forced by hourly surface winds and pressure from global climate model simulations were conducted. The spatial pattern in the long-term trend of annual maximum sea surface height and wave height from the past to the warmer future is predominantly driven by changes in tropical cyclone frequency. In the western North Pacific, future tropical cyclone frequencies are projected to decrease and the annual maximum sea surface height and wave height show a decreasing trend (-20cm/100-yr・-200cm/100-yr). Although very intense tropical cyclones are enhanced in the warmer climate, very extreme storm surge and wave heights do not necessarily increase due to the large natural variability.

 In this study, we examined future changes of explosive cyclones around Japan and their effects on waves along the nation-wide coast using a seamless 150-year high-resolution climate projection from the present to the future climate under warming conditions. The climate projection showed that the number of intense explosive cyclones tends to increase in the future, and the average intensity also increases, and the path tends to shift to the north. In addition, there is a correlation between the trend of the monthly maximum significant wave height and the change in the intensity of explosive cyclones along the coast of Japan and the contribution of explosive cyclones to the annual maximum significant wave height will increase in the future.

 In this study, the effects of global warming on tropical cyclones (TCs) were evaluated by using a slab-ocean model coupled with the Atmospheric Global Climate Model (MRI-AGCM) . The warming conditions for the MRI-AGCM climate simulations were obtained from the latest climate models (CMIP6). We have conducted two types of global climate simulations, typical and newly proposed experiments. The former is time slice experiments and latter is the climate experiment fixed single month condition. Both experiments are conducted on present and future periods, to estimate the effects of climate change impacts on TC characteristics. The slab ocean model which considers sea surface cooling due to strong winds, significantly contributes to a reduction in TC intensity. Moreover, we have confirmed that climate change can reduce the TC frequency and enhance the TC intensity.

 For climate adaptation planning for coastal area, understanding sea level rise on regional-scale is required. We have to assess not only changes in the global averaged sea level but also natural variability in local sea level. In this study, we conducted the analysis of dynamic sea level projection data from the global climate model (MIROC6) and revealed the spatial and temporal characteristics of natural variability in sea level over the North Pacific. We identified representative modes of natural variabilities in sea level including ENSO based on the EOF analysis. The representative EOF modes cause the temporal and local sea level change by more than 10 cm in the low latitude area and the Kuroshio area. In contrast, we found that smaller-scale phenomena dominate the sea-level changes in the Kuroshio extention area than in the North Pacific.

 In this study, we verified a bias correction method for the central pressure of the d4PDF (MRI-AGCM3.2H) typhoon track data assuming the storm surge estimation by the parametric typhoon model. Then, the future changes in the central pressure and return period of typhoons passing through the Pacific coast of Japan are estimated. The results showed that the model bias can be corrected well by applying the quantile mapping method to a low central pressure for each 2.5 latitude division, the typhoon central pressure tends to decrease for a longer return period than 10 years from future experiments, and the range of the SST model prediction uncertainty was up to 20hPa due to differences in future changes of the SST patterns.

 We evaluated the impact of climate change on coastal waves using neural network (NN) model to predict Yorimawari waves at Toyama Bay developed by Masuda et al. and optimized to JRA-55. The large ensemble climate projection data d4PDF is used to estimate the waves to this model. The d4PDF data needs interpolated to JRA-55 grids and bias correction because of the different resolution from JRA-55 and has model bias. Bias correction is used three kinds of bias correction methods to compare the wave estimations. The yearly means of the estimated significant wave height and period with corrected d4PDF in Toyama Bay were consistent with them estimated by JRA-55. The future change of the frequency of high-swell was estimated to decrease. However, the future changes of the maximum wave height and period of the year were estimated to increase slightly.

 Pseudo-global-warming experiments were conducted for typhoon Haiyan using the 3-dimensional atmosphere-ocean difference caused by climate change directly. These experiments aim to understand the future change of typhoon under consistent method in civil engineering. The ocean future change in CMIP5 showed increase of 2.6 – 2.9 ℃ in the surface mixed layer, with little change in the current velocity. The series of experiments were conducted using the coupled atmosphere-ocean-wave model COAWST, and the typhoon in the historical climate reproduced the Best Track well. The typhoons in the future climate showed a decrease in the central pressure of about 34 hPa and an increase in the maximum wind speed of about 12 m/s at the time of fully development, compared to the typhoon in the historical climate. The surge difference obtained by the coupled model corresponds to the suction effect due to the central pressure change. The sea surface temperature increase strengthened the typhoon. The anomaly of sea water temperature associated with the passage of the typhoon were almost the same in the historical and future climate.

 The dynamic direct downscaling simulations by the high-resolution typhoon model and storm surge model for the top 50 typhoon cases are conducted in this study to evaluate the future changes of the L1/L2 storm surges in Ise Bay, Japan. The top 50 cases of extreme typhoons in the past and future climates are selected from the GCM database in the long-term ensemble projections d4PDF. The L2 storm surge under the future climate (its future change) is about 6.49 m (+1.39 m) at the inner part of Ise Bay and about 4.41 m (+0.74 m) at the west coast of Ise Bay. The L2 storm surge is enhanced especially at the inner part of Ise Bay, due to extreme typhoons with higher moving speeds under the future climate. The L1 storm surge under the future climate (its future change) is about 3.94 m (+0.51 m) at the inner part of Ise Bay and about 3.55 m (+1.03 m) at the west coast of Ise Bay. The L1 storm surge is increased more at the west coast of Ise Bay, owing to extreme typhoons with lower moving speeds under the future climate.

 Turbulent mixing in ocean upper layer induces cooling of sea surface temperature (SST) and reductions of heat flux from ocean to typhoon. Sea surface turbulent kinetic energy (TKE) induced by wave dissipation is one of the main source of upper layer turbulence. In this study, two kinds of numerical experiments targeting typhoon Neoguri (2014) have been conducted, hindcast and idealized experiment under fixed atmospheric condition. The results of these experiments show that larger surface TKE flux results in great viscosity in upper layer, which reduces surface current velocity. Surface TKE also accelerates “weak mixing” inside of mixing layer, and then enhances the SST cooling. In addition of the impacts in mixing layer, surface TKE flux stimulates “strong mixing” and wind-induced internal wave which is excited beneath mixing layer. Because of the shallow mixed layer, the strong mixing gives feedback to surface current and SST. These differences due to surface TKE flux also impact on atmospheric condition of typhoon such as sea level pressure through sea surface heat flux. This implies the importance of wavedependent surface TKE for typhoon simulation.

 Towards in-depth understanding of oceanic microplastic contamination as one of the key global environmental problems, Lagrangian particle tracking experiments were conducted to evaluate the transport of microplastics (MPs) derived from the four major rivers that have been known to discharge a huge amount of plastic wastes into the South China Sea (SCS). We carried out two types of the experiment: 1) 2D horizontal tracking of MP particles placed at surface to represent positively buoyant (light) MPs, and 2) 3D tracking of neutrally buoyant MP particles that are passively transported by ambient flow, using the precomputed 3D currents from the climatological HYCOM-ROMS oceanic model. The seasonally varying monsoons in the SCS are found to provoke strong seasonal variability in the river-derived MP transport. It is found that a large portion of MPs especially from the south China coast are transported to the East China Sea in the seasons when the southwesterly monsoon dominates, potentially contributing to the SCS-originated MPs reaching around Japan. The difference of MP density affects their oceanic transport and fates, in which the buoyant MPs are tend to be transported to and accumulated around the nearshore areas due to wind-driven surface shoreward flow.

 At present, there are concerns about changes in the coastal environment due to sea level rise due to global warming all over the world, and the resulting effects are thought to be the disappearance of not only sandy beaches but also the vegetation that inhabits them. In this study, we investigated the effects of sea level rise focusing on beach plants throughout Japan. Eight kinds of beach plants can be extracted, and as a result of predicting the disappearance of vegetation that may be affected by the rise in sea level, the future vegetation area even considering the transition of vegetation with a large disappearance rate. Was found to be likely to decrease. In addition, as a result of examining future appropriate habitat and changes in vegetation area compared with the current situation, the average elevation of the distribution area of Kohamagiku is high in the appropriate habitat, and the habitat suitability area is smaller than that of other beach plants. ing. Regarding the change in vegetation area, the vegetation area of the Leymus mollis-Carex kobomugi group may be lost by about 180,000 m2 compared to the present due to the sea level rise of 1.0 m, and it can be said that it is greatly affected by the disappearance due to the sea level rise.

 In this study, a method for predicting the drift of vessels during typhoons is developed, and the accuracy of predicting the drift of vessels is investigated by calculating the drift using various weather prediction models. The prediction method is based on the domain meteorological model WRF, the Japan Meteorological Agency's GSM (Japan region), and the forecast data of MSM. The results of the prediction calculations showed that, on average, the error of the predicted drifting position from the measured value was about 1 to 4 km, but in the case of a vessel at anchor, the prediction accuracy of the drifting wake and the drifting point was higher in the case with a higher prediction accuracy of the wind field. However, in order to improve the accuracy of predicting the anchorage of a vessel during a typhoon, it is necessary to take into account various uncertainty factors, such as the anchor chain, waves, and the use of the main engine, in order to improve the calculation accuracy, such as the conditions for the start of drifting.

 Beach trench surveys were conducted on a sand gravel beach protected by detached breakwaters in Kochi Coast. Alongshore and cross-shore trenches were excavated with a total length of 150m. The trench faces were recorded by a RTK-UAV and compared with topography data. A clear boundary between sand and gravel layers was observed at a level of 2m above sea level. The bottom gravel layers were considered to be developed by swash waves at the construction of detached breakwaters. The beach was found to be eroded by several high waves in last 15 years. The erosion level, however, was found to be up to the level of sand-gravel boundary, which implied that the sand gravel beach protected by detached breakwaters was resistant to erosion owing to the presence of buried gravel layers.

 The authors have developed a new method to measure the ground height in real time using optical fibers, and have conducted measurements from October 2019 to March 2021. The validity of the measure results were confirmed in the previous paper. In this study, we focus on Typhoon No. 9 and No. 10 which occurred in September 2020, and examined the geomorphological response of delta using the filed observation data.

 Typhoon No. 9 and Typhoon No. 10 were characterized by swell waves and wind waves, respectively. During the passage of both typhoons, the ground level was decresed by 4cm to 8cm, indicating that the high concentration was maintained even when the ground level was decreased. The correspondence between the brightness of the seawater and the L-value of turbidity was also good.

 The DLVO theory using the potential energy around the deposited sediment particles is defined by the sum of the intermolecular energy of the particles and the electrostatic repulsion energy. Thus, we hypothesised that if the surface potential of the bottom sediment particles is measured, the shear stress of the bottom sediment can be evaluated. In this study, we proposed a method to measure the surface potential of bottom sediment particles by utilizing the characteristics of metal electrodes and carbon electrode with semiconductor characteristics. It was experimentally confirmed that the surface potential of the sediment particles generated in the sedimentation process, consolidation and winding up can be directly measured.

 The critical shear stress of the sediment is an important parameter that determines the transportation of the sediment. However, it is difficult to carry out experiments on a large number of samples because a large-scale experimental device and a large number of samples are required to measure the critical shear stress of the sediment. On the other hand, in the measurement using a rotational viscometer, the result obtained is the yield value. However, the experimental equipment is not large, the operation is simple, and the amount of sample required for the experiment is about 200 mL. In this study, in order to establish a simple method for measuring the critical shear stress of sediment, the relationship between the yield value obtained from the rotational viscometer and the critical shear stress obtained from the circulating water tank were investigated using various type of sediment.

 From the experimental results, it was found that there is a correlation between the yield value and the critical shear stress. Moreover, the relationship between the yield value and the water content of the sediment can be expressed by water retention. From these results, the relationship between the water content of sediment and the critical shear stress can be estimated by measuring the water content after centrifugation, which is an index of water retention.

 Tsunami events in Indonesia from 2010 to 2018 show that people respond differently and were not evacuated uniformly. A standard model of evacuation trigger ranking was developed using nine post-disaster evacuation surveys and shows that there are 6 main tsunami evacuation triggers in Indonesia with seeing other evacuees was the prevalent trigger and gave the high influence. Analyses of videos of human behaviors during the evacuation of the 2018 Palu tsunami revealed that (1) late evacuees who noticed the arrival of a tsunami had a greater influence than early evacuees, (2) influence of evacuation triggers increase gradually as the distances between the trigger sources and persons decrease, and (3) evacuees did not know the evacuation routes. These results have enabled improvement of the recent evacuation model. To obtain satisfactory results, the input parameter should not be constant and adjusted using only questionnaire surveys. Both simulation and footage reveal a problem that Indonesians did not evacuate immediately upon feeling the ground motions, and many people who responded promptly did not successfully complete evacuation. This is a pilot research on evacuation start in Indonesia, thus, more studies are required for tsunami evacuation behaviors. This study provides insights into the behaviors of Indonesians during disaster evacuation and might help determine proper countermeasures.

 Tsunami evacuation route searching tool, which based on superposition diagram of different tsunami inundation areas, was developed for reduction tsunami human damage. Crowd evacuees were simulated to move toward the less-possiblly inundated area, by using the proposed tool. The effect of that tools was confirmed with the crowd evacuation simulation. The study area, located on west side of Izu-peninsula, was selected and Suruga-traph, which is the source of tsunami induced by Tokai Earthquake, distributes in front of coastal line of this area. The tsunami inundation simulations with four tsunami source from Mw 8.1 to Mw 9.1 were conducted. The effects of the proposed tool were: (1) the tsunami damage reduction effect is larger in the case with comparatively higher frequence tsunami. (2) According to the simulatin result of daytime and nighttime case which were different from the evactuatin start time after earthquake occurance event and moving speed on evacuation action, the effects of the tool is small in the cases of the large amount of inundation area, and the evacuation start time is very important for lives saving.

 Cognitive dissonance is one of the factors that inhibit evacuation during a disaster. The purpose of this study is to clarify the impact of cognitive dissonance on evacuation by using data obtained from questionnaire surveys on evacuation behavior at the time of the typhoon No. 21 (Jebi) storm surge disaster in 2018. The results of the questionnaire survey showed that few people actually evacuated despite being aware of the possibility of flooding, indicating that cognitive dissonance may have occurred. In addition, we developed a logit model of the choice of evacuation behavior and conducted a quantitative analysis of the impact of cognitive dissonance on evacuation behavior. The values of the parameters estimated by the analysis indicate that people unreasonably underestimated the probability of death or disaster occurrence. Therefore, cognitive dissonance was found to occur during the actual evacuation.

 As shown in the Great East Japan Earthquake and other tsunami disasters, tsunami-related fire can occur even in tsunami inundation areas. Previous studies developed models to assess tsunami-related fire occurrence risk, using data from an assessment in the target area. A tsunami flow can, however, push flammable materials in the targrt area from other areas, that is an important feature of tsunamis. This study constructed logistic regression models in each of Yamada Town, Otuti Town and Kesennuma City to assess tsunami-related fire occurrence risk, using data from the Great East Japan Earthquake. The models included data that related with building damage in areas surrounding a target area, especially seaward areas, as well as that of the target area. Comparison of the constructed models showed that consideration of building damage in the seaward areas is important to predict the tsunami-related fire occurrence, while necessity of including topographic characteristics such as hills into stastical model is shown to futher improve the prediction accuracy.

 In this study, in order to apply the victim information of the Great East Japan Earthquake provided by the Miyagi Prefectural Police Headquarters to the human casualty function, the relationship between the inundation depth and the ratio of drowning victims based on the victim’s address and the corpse location was examined at the central part of Ishinomaki City, Miyagi Prefecture. As a result, it is clarified that occurrence of human casualties was considered to be related to not only inundation depth but also the land use patterns. In addition, we developed human casualty functions by using GLM and GLMM. As a result, it was clarified that it is important to consider the land use pattern as a variable for predicting human casualties.The fragility model based on the victim’s address examined in this paper can be applied for disaster prevention plan. In contrast, the fragility model based on the corpse location can be applied for lifesaving and search for missing persons when disaster happen.

 The Great East Japan Earthquake of March 11, 2011 caused enormous damage in Japan, and various studies have been conducted on the actual damage and its reduction. Hypothermia in cold weather disaster is discussed as a new issue to be considered in tsunami disaster. In this study, the hypothermia based on the location of the bodies in the zip code area of Miyagi Prefecture using the data of the disaster were examined to investigate the actual damage of hypothermia in the disaster. The analysis of the location information of the victims indicated that they may have developed hypothermia indoors at the evacuation site. The analysis of the flooding situation in the area where the hypothermia victims were found indicated that they may have developed hypothermia due to the wetness of their bodies caused by the exposure to the tsunami. As a new hypothesis of hypothermia risk, a case in which a person whose body was wet by the tsunami develops hypothermia indoors at an evacuation site was presented, and the importance of taking measures against hypothermia at evacuation sites was demonstrated.

 In recent years, the risk of natural disasters is expected to increase due to climate change. However, the impact of sea level rise on tsunami damage and related economic loss is not clear. Therefore, this study made an attempt in evaluating future tsunami damage in Tohoku region as an example. Future earthquake along Japan Trench is selected as tsunami source and combined with sea level rise scenarios. Economic loss was performed by using an economic model with land use data and Input-Output Tables. As a result of estimating the tsunami hazard and risk for the current and multiple sea level rise scenarios, it was found that the total damage in the Tohku region would increase by about 46% for a 1-meter sea level rise. The results of the prefecture-by-prefecture analysis showed that Aomori and Miyagi were particularly vulnerable as well as agricultual sector which found to be the most vulnerable when considering sector by sector. Such findings can suggest for emergency planning for each perfecture and industry against future tsunamis.

 The field survey method used to evaluate ecosystem services for tourism/ recreation and relaxation are limited by a number of challenges, including labor, cost, and the insufficiently of past data accumulation. By focusing on mobile phone network data, we devised an alternative method for estimating the number of visitors and the consumer surplus; this method was tested at the Yokohama Umi-Koen where we conducted a field survey from 2013 to 2014. The spatial resolution of mobile phone network data is 25ha, which is coarser than the area of the Yokohama Umi-Koen (15ha); thus, the number of visitors was overestimated due to impact of the people outside the evaluation are, such as the residential area. To circumvent this problem, we devised a method for estimating the population in the residential area, which reduced the risk of overestimating visitor numbers. As a result, we obtained a closer correlation to the visitor estimates from the field survey. In addition, we verified the relevant amount of mobile phone network data required to estimate the annual number of visitors. With this data, we were also able to estimate the consumer surplus using the zone travel cost method.

 In this study, we examine the monitoring of coastal topography by creating a dataset that combines "shoreline position readings from open satellite images" and "depths and locations obtained from fishing vessels during operation and navigation (fishing vessel big data)." We show the changes in coastal topography along 8 km of coastline from 2017, 2018, and 2019 data collected around the Asaba Coast in Enshu-nada, Shizuoka Prefecture, and explain what can be read from the topographic data and its applicability to monitoring coastal topography. Further, we discuss the method's applicability as a coastal monitoring system that replaces or is used with surveying.

 Distribution of coastal sediments needs to be known for effective coastal management periodicals. In this study, the classification model of coastal sediments using the convolution neural network, which was supposed in the last year by the author, was improved by adopting the Gaussian blur to pictures of training the model. The results of training of the model showed that the accuracy of training and test improved compared with the result of training with original pictures. Implementation of model for aerial pictures indicated the sand area were classified with not enough accuracy. This result shows Gaussian blur decrease features of sand in training pictures. In the results implemented to aerial pictures in the gravel area, increasing of strength of Gaussian blur provided improvement of classification accuracy of gravels. However, some area of gravels still classified as sand in all models. In order to improve of model additionally, it was estimated that the numbers of pictures for training should be increased.

 In order to monitor the particle size distribution of sandy beach with high frequency, high density, wide area and long term, we are trying to develop an image analysis method based on autocorrelation function for sediment images as an alternative method to sieve analysis. In this study, in order to understand the effect of sandy beach characteristics on the measurement accuracy, field surveys were conducted at 133 sandy beaches, and samples were collected for image analysis. In the range of the median particle size up to 2 mm, the accuracy of the median particle size measurement was shown to be equivalent to that of sieve analysis. On the other hand, the accuracy tended to decrease in the case of fine-grained sediment or when the contrast of the gray-scale image was small. It was shown that the shoreline retreat when the sea level rises in the future can be equivalently evaluated by the image analysis method when the retreat is up to about 20 m.

 In this study, shorelines from four Japanese beaches were extracted from 85 optical satellite images to evaluate the feasibility of two automated edge detection methods. The shorelines were successfully detected on both sandy beaches and gravel coasts. The method based on deep learing showed a high extraction rate and accuracy as compared with the unsupervised method based on the Canny edge filter and Otsu thresholding. The accuracies of the extracted shoreline were roughly correlated to the cloud coverage. Yearly composite images were generated to remove the effect of cloud, and showed a high extraction rate and accuracy when analysed by the same methods. Some considerations for utilizing these methods on shoreline monitoring were discussed.

 Coastal wind speeds retrieved from Sentinel-1 synthetic aperture radar (SAR) with four types of wind directions (in situ, Japan Meteorological Agency mesoscale analysis (MANAL), Sentinel-1 Level 2 Ocean (OCN) product and extracted wind streaks from SAR images with 2D FFT) are compared to clarify a suitable wind direction. The result shows that in situ, MANAL and OCN wind directions can retrieve SAR wind speeds with a root mean square error (RMSE) of around 25% though the variation width of biases with the MANAL wind direction is slightly large. It is also found that the mean absolute error of the SAR wind direction extracted by 2D FFT is larger than the others especially when the wind speed is low. The error of wind direction is one of the reasons for a large RMSE (30.5%) of the SAR wind speed. It is concluded that the in-situ wind direction is the most reliable, but the OCN and the MANAL wind directions can also be used for the SAR wind speed retrieval.