The concept of a digital twin [1] , which involves analyzing information obtained from the analysis of detailed and large-scale digital models replicating real-world entities and providing feedback to real-world predictions and control, is an essential technological concept aimed at realization in various fields [2] . In the field of civil engineering responsible for managing urban societies and disaster prevention, modeling entire urban cities based on the concept of digital twin is positioned as a fundamental initiative to enhance the understanding the state of infrastructures and disaster situations [3] , [4] , [5] .

Lagrangian persistence method in nowcasting is a highly effective method for short-term weather forecasting. However, its performance is not as robust in long-term forecasting or situations of rapid weather changes due to the difficulty in analyzing the intensity variation of meteorological physical quantities. To address this shortcoming, in this study, we incorporated Dynamic Mode Decomposition into the Lagrangian persistence method. Specifically, we proposed a Coordinate-Transformed Dynamic Mode Decomposition (CT-DMD) model by integrating an Optical Flow model with a Dynamic Mode Decomposition model, providing an effective method for analyzing the intensity variation of meteorological physical quantities in the Lagrangian persistence method. The integration of the Optical Flow model and the Dynamic Mode Decomposition model involves the transformation of data between Eulerian and Lagrangian coordinate systems. The CT-DMD model was evaluated using radar-observed rainfall data from the Kanto region of Japan, with the Rainymotion model used as a benchmark. When the lead time was 5 min, 22.22% of the subsets in the experimental dataset showed that the CT-DMD model had a higher forecast accuracy compared to the Rainymotion model. When the lead time was 25 min, 88.89% of the subsets in the experimental dataset showed that the CT-DMD model had a higher forecast accuracy compared to the Rainymotion model. The accuracy advantage of the CT-DMD model became apparent after a lead time of 15 min and became increasingly significant as the lead time increased. The results demonstrated the validity of the CT-DMD model.

The utilization of multimodal analysis techniques, combining satellite imagery with terrain information, has gained prominence in flood detection. This study focuses on the inclusion of elevation data into the Sen1floods11 dataset, an open dataset for flood damage detection, to investigate the influence of terrain information on flood detection tasks. Among the considered terrain information, the inclusion of elevation data resulted in a bias of overestimating the presence of water in relatively low-lying areas, without contributing to accuracy improvement. However, the utilization of slope information, derived from differentiating the elevation data, mitigated such bias and yielded a slight improvement in accuracy. This finding aligns with the utilization of derivatives in physical equations describing flood flow, suggesting the explicit incorporation of physics-based principles, such as the flow of water based on slope, to enhance model accuracy in future research endeavors.

Floods are one of the most severe natural catastrophes and therefore emergency response operations are crucial in order to save lifes. These operations require information about flooded areas so that rescue missions can precisely and efficiently use their available resources. This requires a quick automated procedure which is able to identify these regions from remote sensing images. To achieve this goal we utilize machine learning and apply our method on the Sen1Floods11 dataset. Our main contribution lies in the fusion of Digital Elevation Maps (DEMs) with Satellite data. We investigate the effect of several different combinations of processing methods of DEMs, such as depression filling, deriving slope and curvature or flow metrics. In total 44 different experiments have been performed where our best performing combination outperformed the benchmark in terms of mean IoU. Lastly, we also publish our code for downloading and processing DEMS as well as running our experiments.

With the development of big data and advances in computing capability, data-driven models have shown significant potential in various fields. In this regard, several studies have focused on using meteorological big data for short-term rainfall predictions. However, limited efforts have been devoted toward elucidating the physical nature of meteorological phenomena. To address this, we constructed a physics-informed data-driven model for short-term precipitation predictions using meteorological big data. The spatial movement of precipitation was predicted via the optical flow method, and the substantial development and decay of precipitation was predicted using Koopman operator analysis. We verified the accuracy of the proposed model based on the radar-observed precipitation data for 2015 and 2017 in Japan. The results reveal that the model can effectively analyze changes in the physical quantities and dynamic systems of atmospheric fluids.

Herein, we propose a novel scheme for detecting earthquake-damaged buildings from optical satellite imageries. The scheme comprises two steps: first, using the information of the positions and shapes of the buildings in a GIS database, we identify the photographic scope of each residence in wide-area photographic imagery and extract small photographic fragments at the level of individual structures. Second, using a classifier to determine if these individual fragments represent collapsed structures, we assess the damage of the residential structures in the affected area. Furthermore, we verify the effectiveness of applying the recent machine learning techniques to improve the performance of the classifier. To utilize images captured before and immediately after the earthquake, we apply spatio-temporal convolution neural network that is regarded as a generalized method of image subtraction. Additionally, we integrate the structural attributes comprising structural age and structural materials with the satellite images using a multimodal learning structure. The effectiveness of using the aforementioned techniques is discussed using a dataset constructed from a satellite imagery of the affected area in 2016 Kumamoto Earthquake Japan taken by Spot 6 and 7.

We propose to use bootstrap samples in designing an extended quadratic classifier. The proposed method is to generate bootstrap samples which have much information about distributions, and to optimise the extended quadratic classifier so that the error estimated by using bootstrap samples is minimized. The performance of the proposed classifier is demonstrated on real data.

This report describes the required elements for a next-generation document reader. These results are derived from the analysis of the user requirements. The next-generation document reader will have a high degree of adaptability to a wide range of applications and specifications. Key requirements for the document reader are accurate character recognition, accurate layout analysis, portability, and adaptability. The prototype is developed to test the next-generation document reader's functions.