SYSatoru Yamaguchi

Satoru Yamaguchi

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article

Previous Organizations / Institutes

The organizations / institutes from the scholar's other articles in the past ten years


Satoru Yamaguchi has a total of 181 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    Surface energy and mass balance variations near the equilibrium line altitude on Qaanaaq Ice Cap, northwestern Greenland, 2012–2022

    by Motoshi Nishimura, Teruo Aoki, Masashi Niwano, Motomu Oyama, Sumito Matoba, Tomonori Tanikawa, Satoru Yamaguchi, Rigen Shimada, Takumi Suzuki

    Rapid warming of the Arctic has accelerated the mass loss of snow and ice. The glacier equilibrium line altitude (ELA) is a suitable location for monitoring the surface energy balance (SEB) and surface mass balance (SMB). This study presents SEB and SMB variations during 2012–2022 around the ELA on the Qaanaaq Ice Cap in northwest Greenland. The SMB was analyzed using SEB analysis and the surface height change method (SHM) with in situ meteorological data. Results revealed that the most significant surface melt occurred in the summers of 2019 (1058 and 1191 mm w.e. a −1 ) and 2015 (974 and 731 mm w.e. a −1 ) for the SEB and SHM, respectively. Additionally, the primary energy contributing to the surface melt in 2014/15 was net shortwave radiation caused by snow/ice albedo feedback, whereas in 2018/19, it was not only shortwave radiation but also increased downward longwave radiation from warm cloud cover. Furthermore, the cumulative SMB over the decade (2012–2022) is slightly positive, with a value of +0.26 and +0.28 m w.e., for the SEB and SHM, respectively. The results on the amplification of the SEB and SMB around the ELA provide valuable insights into monitoring environmental changes in the Arctic.

  • Open Access

    Educating for Climate-Driven Geohazard Mitigation and Management – Experience from the GEOMME International Partnership in Norway, South Korea, and Japan

    by Graham Lewis Gilbert, Vittoria Capobianco, Luca Piciullo, Dieter Issler, Satoru Yamaguchi, Yoichi Ito, Takahiro Tanabe, Ryoko Nishii, Hirofumi Niiya, Tae-Hyuk Kwon, et al.

    The GEOMME partnership for Geohazard Mitigation, Management, and Education is an international collaboration between research institutes and universities in Norway, Japan, and South Korea. The partnership aim has been to enhance resilience to climate-driven geohazards through collaborative research and education. The project is funded by the Research Council of Norway (pnr 322469) with a duration from September 2021 to December 2026. A main objective of GEOMME has been to initiate collaborative activities that improve the collective ability of the partner countries and institutes to respond to current and emerging disaster risk management challenges to climate-driven geohazards through knowledge exchange and research-based education.Over the past four-years, the GEOMME partners have developed and hosted four education packages focused on different aspects of climate-driven hazard and risk management. The education packages were structured as modular courses, addressing: (i) geohazards and risk in a changing climate (hosted in Tromsø, Norway in 2022), (ii) large-scale hazard and risk assessment (hosted in Niigata, Japan in 2023), (iii) monitoring, modelling, and early warning (hosted in Daejeon, South Korea in 2024), and (iv) sustainable and nature-based mitigation strategies (hosted in Florence, Italy in 2025).Each education package consisted of a digital pre-study module followed by a two-week intensive course. The pre-study modules were used as a level-setting tool for participants with different academic backgrounds prior to attending the courses. The intensive courses combined field- and research-based training and scenario-driven learning. A key objective of the in-person activities was to bring together an international group of students, practitioners, researchers, and educators.The aim of this contribution is to: (i) share the developed educational material and present the GEOMME partnership as a potential model for international, research-based geoscience education and (ii) reflect on key lessons learned related to interdisciplinary teaching and the transferability of this approach to other contexts.

  • Impacts of the December 2022 Heavy Snowfall on Tree Fall and Power Outages in Sado City, Japan

    by Sojiro Sunako, Satoru Yamaguchi, Kazuki Nakamura, Kengo Sato, Hiroki Motoyoshi

    On December 18, 2022, heavy snowfall in Sado City, Niigata Prefecture, caused widespread tree fall and bamboo collapse, and subsequent power outages. The total number of households affected by the outage reached 17,510 during the event, and restoration of the power supply required approximately 11 days. We conducted field surveys to investigate the characteristics of damage to trees and bamboo, as well as the associated power outages. This survey was conducted using both visual inspections and a vehicle equipped with an artificial intelligence-based road surface assessment system that employed a smartphone camera to document the damage to trees and bamboo and condition of road surface. Additionally, meteorological data and power outage records were analyzed to clarify the detailed weather conditions during the outage events and consider their relationship with the field survey. The results revealed that stem breakage of trees predominantly occurred in mountainous areas at relatively high elevation, whereas bamboo collapse was primarily observed in lowland areas. Analyzing meteorological data and outage records indicated that persistent strong northwesterly winds and intermittent snowfall contributed to snow accretion on trees, leading to uneven snow loading and increased susceptibility to wind damage. These conditions likely triggered widespread damage to trees and bamboo and ultimately resulted in power outages. Furthermore, the presence of dense bamboo stands and dead bamboo adjacent to power lines, which are particularly vulnerable to snow and wind damage, was considered to have contributed to the extensive power outages observed in the area.

  • Open Access

    Learning to filter: snow data assimilation using a Long Short-Term Memory network

    by Giulia Blandini, Francesco Avanzi, Lorenzo Campo, Simone Gabellani, Kristoffer Aalstad, Manuela Girotto, Satoru Yamaguchi, Hiroyuki Hirashima, Luca Ferraris

    Trustworthy estimates of snow water equivalent and snow depth are essential for water resource management in snow-dominated regions. While ensemble-based data assimilation techniques, such as the Ensemble Kalman Filter (EnKF), are commonly used in this context to combine model predictions with observations therefore to improve model performance, these ensemble methods are computationally demanding and thus face significant challenges when integrated into time-sensitive operational workflows. To address this challenge, we present a novel approach for data assimilation in snow hydrology by utilizing Long Short-Term Memory (LSTM) networks. By leveraging data from 7 diverse study sites across the world to train the algorithm on the output of an EnKF, the proposed framework aims to further unlock the use of data assimilation in snow hydrology by balancing computational efficiency and complexity. We found that a LSTM-based data assimilation framework achieves comparable performance to state estimation based on an EnKF in improving open-loop estimates with only a small performance drop in terms of RMSE for snow water equivalent (+6 mm on average) and snow depth (+6 cm), respectively. All but 2 out of 14 site-specific-LSTM configurations improved on the Open Loop estimates. The inclusion of a memory component further enhanced LSTM stability and performance, particularly in situations of data sparsity. When trained on long datasets (25 years), this LSTM data assimilation approach also showed promising spatial transferability, with less than a 20 % reduction in accuracy for snow water equivalent and snow depth estimation. Once trained, the framework is computationally efficient, achieving a 70 % reduction in computational time compared to a parallelized EnKF. Training this new data assimilation approach on data from multiple sites showed that its performance is robust across various climate regimes, during dry and average water-year types, with only a limited drop in performance compared to the EnKF (+6 mm RMSE for SWE and +18 cm RMSE for snow depth). This work paves the way for the use of deep learning for data assimilation in snow hydrology and provides novel insights into efficient, scalable, and less computationally demanding modeling framework for operational applications.


  • Open Access

    Reply on RC2

    by Giulia Blandini, Francesco Avanzi, Lorenzo Campo, Simone Gabellani, Kristoffer Aalstad, Manuela Girotto, Satoru Yamaguchi, Hiroyuki Hirashima, Luca Ferraris

    Trustworthy estimates of snow water equivalent and snow depth are essential for water resource management in snow-dominated regions. While ensemble-based data assimilation techniques, such as the Ensemble Kalman Filter (EnKF), are commonly used in this context to combine model predictions with observations therefore to improve model performance, these ensemble methods are computationally demanding and thus face significant challenges when integrated into time-sensitive operational workflows. To address this challenge, we present a novel approach for data assimilation in snow hydrology by utilizing Long Short-Term Memory (LSTM) networks. By leveraging data from 7 diverse study sites across the world to train the algorithm on the output of an EnKF, the proposed framework aims to further unlock the use of data assimilation in snow hydrology by balancing computational efficiency and complexity. We found that a LSTM-based data assimilation framework achieves comparable performance to state estimation based on an EnKF in improving open-loop estimates with only a small performance drop in terms of RMSE for snow water equivalent (+ 6 mm on average) and snow depth (+ 6 cm), respectively. All but 2 out of 14 LSTM site specific configurations improved on the Open Loop estimates. The inclusion of a memory component further enhanced LSTM stability and performance, particularly in situations of data sparsity. When trained on long datasets (25 years), this LSTM data assimilation approach also showed promising spatial transferability, with less than a 20 % reduction in accuracy for snow water equivalent and snow depth estimation. Once trained, the framework is computationally efficient, achieving a 70 % reduction in computational time compared to a parallelized EnKF. Training this new data assimilation approach on data from multiple sites showed that its performance is robust across various climate regimes, during dry and average water-year types, with only a limited drop in performance compared to the EnKF (+6 mm RMSE for SWE and +18 cm RMSE for snow depth). This work paves the way for the use of deep learning for data assimilation in snow hydrology and provides novel insights into efficient, scalable, and less computationally demanding modeling framework for operational applications.

  • Open Access

    Reply on RC1

    by Giulia Blandini, Francesco Avanzi, Lorenzo Campo, Simone Gabellani, Kristoffer Aalstad, Manuela Girotto, Satoru Yamaguchi, Hiroyuki Hirashima, Luca Ferraris

    Trustworthy estimates of snow water equivalent and snow depth are essential for water resource management in snow-dominated regions. While ensemble-based data assimilation techniques, such as the Ensemble Kalman Filter (EnKF), are commonly used in this context to combine model predictions with observations therefore to improve model performance, these ensemble methods are computationally demanding and thus face significant challenges when integrated into time-sensitive operational workflows. To address this challenge, we present a novel approach for data assimilation in snow hydrology by utilizing Long Short-Term Memory (LSTM) networks. By leveraging data from 7 diverse study sites across the world to train the algorithm on the output of an EnKF, the proposed framework aims to further unlock the use of data assimilation in snow hydrology by balancing computational efficiency and complexity. We found that a LSTM-based data assimilation framework achieves comparable performance to state estimation based on an EnKF in improving open-loop estimates with only a small performance drop in terms of RMSE for snow water equivalent (+ 6 mm on average) and snow depth (+ 6 cm), respectively. All but 2 out of 14 LSTM site specific configurations improved on the Open Loop estimates. The inclusion of a memory component further enhanced LSTM stability and performance, particularly in situations of data sparsity. When trained on long datasets (25 years), this LSTM data assimilation approach also showed promising spatial transferability, with less than a 20 % reduction in accuracy for snow water equivalent and snow depth estimation. Once trained, the framework is computationally efficient, achieving a 70 % reduction in computational time compared to a parallelized EnKF. Training this new data assimilation approach on data from multiple sites showed that its performance is robust across various climate regimes, during dry and average water-year types, with only a limited drop in performance compared to the EnKF (+6 mm RMSE for SWE and +18 cm RMSE for snow depth). This work paves the way for the use of deep learning for data assimilation in snow hydrology and provides novel insights into efficient, scalable, and less computationally demanding modeling framework for operational applications.

  • Open Access

    Comment on egusphere-2025-423

    by Giulia Blandini, Francesco Avanzi, Lorenzo Campo, Simone Gabellani, Kristoffer Aalstad, Manuela Girotto, Satoru Yamaguchi, Hiroyuki Hirashima, Luca Ferraris

    Trustworthy estimates of snow water equivalent and snow depth are essential for water resource management in snow-dominated regions. While ensemble-based data assimilation techniques, such as the Ensemble Kalman Filter (EnKF), are commonly used in this context to combine model predictions with observations therefore to improve model performance, these ensemble methods are computationally demanding and thus face significant challenges when integrated into time-sensitive operational workflows. To address this challenge, we present a novel approach for data assimilation in snow hydrology by utilizing Long Short-Term Memory (LSTM) networks. By leveraging data from 7 diverse study sites across the world to train the algorithm on the output of an EnKF, the proposed framework aims to further unlock the use of data assimilation in snow hydrology by balancing computational efficiency and complexity. We found that a LSTM-based data assimilation framework achieves comparable performance to state estimation based on an EnKF in improving open-loop estimates with only a small performance drop in terms of RMSE for snow water equivalent (+ 6 mm on average) and snow depth (+ 6 cm), respectively. All but 2 out of 14 LSTM site specific configurations improved on the Open Loop estimates. The inclusion of a memory component further enhanced LSTM stability and performance, particularly in situations of data sparsity. When trained on long datasets (25 years), this LSTM data assimilation approach also showed promising spatial transferability, with less than a 20 % reduction in accuracy for snow water equivalent and snow depth estimation. Once trained, the framework is computationally efficient, achieving a 70 % reduction in computational time compared to a parallelized EnKF. Training this new data assimilation approach on data from multiple sites showed that its performance is robust across various climate regimes, during dry and average water-year types, with only a limited drop in performance compared to the EnKF (+6 mm RMSE for SWE and +18 cm RMSE for snow depth). This work paves the way for the use of deep learning for data assimilation in snow hydrology and provides novel insights into efficient, scalable, and less computationally demanding modeling framework for operational applications.

  • Open Access

    Understanding snow conditions in Japanese Mountains through a quarter century of insights

    by Sojiro Sunako, Satoru Yamaguchi, Yoichi Ito, Katsuya Yamashita, Hayato Arakawa, Masaki Nemoto

    Japan features a wide variety of snow environments, with significant differences in snow distribution based on region and elevation. The effects of ongoing climate change are expected to alter snow conditions significantly, highlighting the need for a comprehensive understanding of these changes and their implications.Meteorological observation sites managed by the Japan Meteorological Agency are predominantly concentrated in low-altitude areas, with only a limited number in high-altitude regions. Consequently, the lack of long-term and spatially extensive observational data has hindered the quantitative understanding of snow conditions in Japan's mountainous regions. To address this gap, the Snow and Ice Research Center (SIRC) of the National Research Institute for Earth Science and Disaster Resilience has operated a Snow and Weather Observation Network for over 25 years. This network spans a wide area, from northern to western Japan, monitoring meteorological and snow conditions fluctuations at high elevations. This study utilizes the long-term data collected through the network to analyze trends in snow conditions across Japan's mountainous regions. Particular emphasis is placed on examining the variation characteristics of maximum snow depth and maximum snow water equivalent, including their differences, elevation dependencies, and contrasts between mountainous and flat areas.Our results indicate that although the trends in maximum snow depth and maximum snow water equivalent are generally consistent, the variability in maximum snow water equivalent is greater than that of maximum snow depth. Additionally, the variation in maximum snow depth shows a distinct dependence on elevation, with different trends observed in mountainous and flat regions. These findings enhance our understanding of the effects of climate change on Japan's snow environments and provide essential insights for improving future projections of snow conditions, as well as for developing strategies to mitigate snow-related disasters.

  • Open Access

    Quality-controlled meteorological datasets from SIGMA automatic weather stations in northwest Greenland

    by Motoshi Nishimura, Teruo Aoki, Masashi Niwano, Sumito Matoba, Tomonori Tanikawa, Tetsuhide Yamasaki, Satoru Yamaguchi, Koji Fujita

    In situ meteorological data are essential for a better understanding of the ongoing environmental changes in the Arctic. In order to increase the scientific value of discussions on understanding the actual state of environmental change in a given area, it is necessary to appropriately remove the anomalous values recorded due to external factors resulting from low temperature and icing. Here we present methods for quality control (QC) of meteorological observation datasets from two automatic weather stations in northwest Greenland, where drastic glaciological and meteorological environmental changes have occurred. The stations were installed in the accumulation area of the Greenland Ice Sheet (SIGMA-A site, 1490 m a.s.l.) and near the equilibrium line of the Qaanaaq Ice Cap (SIGMA-B site, 944 m a.s.l.). We describe the two-step sequence of QC procedures we used to produce increasingly reliable data sets by masking erroneous records. This method was developed for the climatic conditions of Greenland, however, it is designed to be as universally applicable as possible, with a basis in meteorology and glaciology, and with the intention of removing the subjectivity of the person performing the QC. The QC is divided into two processes: Initial Control and Secondary Control. Initial Control removes values that violate physical laws and also serves as a preliminary process to improve the accuracy of Secondary Control. Secondary Control removes abnormal values using stricter statistical criteria than Initial Control. As a result of this two-step process, controlled by scientifically objective criteria, we were able to successfully remove erroneous data sets and greatly reduce the time required for QC. In addition, by using a generally applicable process, we were able to successfully establish an algorithm that could be applied to multiple sites. The data sets from both the SIGMA-A and SIGMA-B sites were classified into three levels (Level 1.1 to Level 1.3) according to the stage of data processing. Level 1.1 is the so-called raw data, in which the data for the period when the logger was stopped are masked (processed to flag them as missing or abnormal), the so-called raw data. Level 1.2 and Level 1.3 are datasets to which Initial Control and Secondary Control have been applied to the Level 1.1 and Level 1.2 datasets, respectively, and the Level 1.3 dataset is a dataset from which all abnormal values have been removed. These datasets have been archived in the Arctic Data Archive System (ADS) operated by the National Institute of Polar Research in Japan (e.g., Level 1.3 dataset: SIGMA-A - https://doi.org/10.17592/001.2022041303 and SIGMA-B - https://doi.org/10.17592/001.2022041306).

  • Open Access

    Learning to filter: Snow data assimilation using a Long Short-Term Memory network

    by Giulia Blandini, Francesco Avanzi, Lorenzo Campo, Simone Gabellani, Kristoffer Aalstad, Manuela Girotto, Satoru Yamaguchi, Hiroyuki Hirashima, Luca Ferraris

    In snow-dominated regions, today’s snow is tomorrow’s water, making reliable estimates of snow water equivalent (SWE) and snow depth crucial for water resource management. In this context, data assimilation is a powerful tool to optimally combine models and measurements, enhancing accuracy and reliability. Ensemble-based techniques like the Ensemble Kalman Filter (EnKF) and Particle Filter (PF) are often used but their deployment in real-time applications can make it challenging to ensure timely and accurate results. To address these challenges, we propose an innovative data assimilation framework for snow hydrology that leverages Long Short-Term Memory (LSTM) networks. Using data from seven diverse study sites across the Northern Hemisphere, our framework is trained on the outputs of an EnKF, persuing a balance between computational efficiency and model complexity to advance data assimilation applications in snow hydrology. This LSTM-based framework achieves performance comparable to the EnKF in improving open-loop estimates, with only minor increases in root-mean-square error (RMSE): +6 mm for SWE and +6 cm for snow depth on average. Adding a memory component enhances stability and accuracy, especially under sparse data conditions. When trained on long-term datasets spanning 25 years, the LSTM framework demonstrated promising spatial transferability, with accuracy reductions of less than 20% for snow water equivalent and snow depth estimation. After training, the LSTM approach significantly outperformed a parallelized EnKF in computational efficiency, reducing runtime by 70% while maintaining comparable accuracy. Training on multi-site data further ensured robust performance across diverse climate regimes and during both dry and average water years, with a modest RMSE increase compared to the EnKF (+6 mm for SWE and +18 cm for snow depth). By combining the strengths of traditional ensemble methods and modern machine learning, this framework offers a scalable, computationally efficient, and reliable alternative for operational snow hydrology data assimilation.