
Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown
Published: April 19, 2024
Language: Английский
Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown
Published: April 19, 2024
Language: Английский
International Journal of Disaster Risk Reduction, Journal Year: 2023, Volume and Issue: 93, P. 103755 - 103755
Published: May 24, 2023
Language: Английский
Citations
5EarthArXiv (California Digital Library), Journal Year: 2023, Volume and Issue: unknown
Published: April 5, 2023
Rainfall nowcasting provides short-term, high-resolution information on the location, intensity, and timing of rainfall, which is crucial for weather forecasting, flood warning, emergency response. This can help people organizations make informed decisions to mitigate impact severe events reduce risk damage loss life. There are many attempts at tackling problem hand, whether it be numerical models or statistical that also comprise deep neural networks. Even though nowcast quite accurate nowadays has a saturated literature, current approaches mostly focus improving performance while computational burden keeps increasing. In this study, we propose EfficientRainNet, convolutional network architecture based mobile inverted residual linear bottleneck blocks with few alterations. We show EfficientRainNet able produce comparable results those encoder-decoder GRUs only fraction trainable parameters over radar rainfall dataset State Iowa. Also, most part, performs better than baselines using persistence- optical flow-based nowcasting, along another efficiency-focused architecture, Small Attention UNet.
Language: Английский
Citations
4EarthArXiv (California Digital Library), Journal Year: 2023, Volume and Issue: unknown
Published: Aug. 3, 2023
Transportation systems can be significantly affected by flooding, leading to physical damage and subsequent adverse impacts such as increased travel distance essential services. Even though flooding is a frequently recurring phenomenon that affect thousands of people per event, there are limited accessible online tools available for analyzing visualizing flood risk supporting decisions on routing emergency planning, response. Existing generally based complicated models not easily non-expert users. Therefore, it critical have efficient real-time communication decision transportation networks various stakeholders, including the public, minimize those groups. This paper presents web application uses graph network methods latest technologies standards assist in describing events terms operational constraints provide analytical support mobility mitigation during these events. The framework designed user-friendly, enabling users access information about road status, shortest paths amenities, location-allocation, service coverage. study area includes following two communities State Iowa, Cedar Rapids Charles City, which were used test application's functionality explore outcomes. Our research demonstrates bridge operation, from locations arbitrary point-to-point routing, planning facility placement, accessibility. introduced capable solving complex flood-related tasks providing an understandable representation vulnerability, enhancing strategies. this provides valuable tool stakeholders make informed
Language: Английский
Citations
4Water Resources Research, Journal Year: 2023, Volume and Issue: 59(9)
Published: Aug. 21, 2023
Abstract The growing number of flood events in urban areas is causing far‐reaching economic losses and social disruptions. To implement suitable risk management strategies measures for prevention mitigation, it paramount to adequately assess economy‐wide associated with potential events. assessment damage the USA relies significantly on Federal Emergency Management Agency (FEMA) Hazus Flood Model. A primary requirement reduce uncertainty empirically gather user input data. Consequently, there a need integrate publicly accessible business data accurate evaluation activities. We present novel approach estimating both direct activities during total impact that arise from these First, we FEMA Model Reference Solutions business‐level estimate reduction (i.e., losses) following 100‐year event New Hanover County, North Carolina. Second, broader activity have industry output, employment, value‐added household income using regional computational general equilibrium model. advantage our easy replicability loss estimation process US cities regions interest, which, when used follow‐up study, leads Policy makers relying solely physical damages would, by ignoring indirect losses, underestimate benefits investing mitigation protection.
Language: Английский
Citations
4Research Square (Research Square), Journal Year: 2024, Volume and Issue: unknown
Published: April 19, 2024
Language: Английский
Citations
1