Related references
Note: Only part of the references are listed.
Article
Environmental Sciences
Kai Mao et al.
Summary: In this study, an information spatial-temporal extension (ISTE) algorithm based on deep neural networks is proposed to predict ocean temperature by fusing satellite remote sensing SST data, ship survey observations data, and historical data. Experimental results demonstrate that the ISTE algorithm outperforms linear regression analysis-based prediction, with high coefficient of determination (0.9936) and low root mean squared errors (around 0.7 degrees C) compared to Argo observation data. Therefore, the ISTE algorithm driven by satellite remote sensing SST can serve as an effective approach for shipborne predictions of ocean temperature.
Article
Environmental Sciences
Siyun Hou et al.
Summary: Sea surface temperature is a crucial factor affecting global climate and marine activities. Existing approaches for predicting different temporal scales train separate models, which is inefficient and cannot utilize correlations. This study proposes a unified spatio-temporal model that can simultaneously predict SST at different scales.
Article
Computer Science, Artificial Intelligence
Yunbo Wang et al.
IEEE TRANSACTIONS ON PATTERN ANALYSIS AND MACHINE INTELLIGENCE
(2022)
Article
Environmental Sciences
Nengli Sun et al.
Summary: Monitoring and predicting sea temperature is crucial for understanding the likelihood of ocean-related events. This study uses a 3D U-Net model to predict the subsurface temperature of the Pacific Ocean and adjacent oceans, showing that this method can provide more accurate predictions compared to previous methods.
Article
Environmental Sciences
Nengli Sun et al.
Summary: This paper proposes a 3D-convolutional long short-term memory (ConvLSTM)-based model for three-dimensional gridded radar echo extrapolation. The model shows better overall and longer-term performance for storms with high reflectivity values and can facilitate early warning regarding impending severe storms.
Article
Meteorology & Atmospheric Sciences
Sandy Chkeir et al.
Summary: Predicting extreme weather events in a short time period and localized areas is challenging. This study develops a machine learning model to nowcast rain and wind speed in the area of Malpensa airport by merging different datasets, with the aim of providing reusable predictions in other locations. The results show that the Long Short-Term Memory Encoder Decoder approach is well suited for nowcasting meteorological variables.
ATMOSPHERIC RESEARCH
(2022)
Proceedings Paper
Computer Science, Artificial Intelligence
Zhangyang Gao et al.
Summary: This paper proposes SimVP, a simple and efficient video prediction model that is built upon CNN and trained by MSE loss. Despite its simplicity, SimVP achieves state-of-the-art performance on multiple benchmark datasets. Extended experiments demonstrate its strong generalization and scalability, with significantly reduced training cost.
2022 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION (CVPR 2022)
(2022)
Article
Geosciences, Multidisciplinary
Xiang Pan et al.
Summary: A novel deep-learning (DL) model, FURENet, is designed to improve nowcasting of convective storms by extracting information from multiple input variables. By incorporating polarimetric radar variables K-DP and Z(DR), the model shows enhanced predictive accuracy. Quantitative statistical evaluation demonstrates that using FURENet, K-DP, and Z(DR) synergistically improve nowcasting skills for convective storms at lead times of 30 and 60 minutes.
GEOPHYSICAL RESEARCH LETTERS
(2021)
Article
Multidisciplinary Sciences
Suman Ravuri et al.
Summary: Advanced nowcasting methods using deep generative models with radar observations can provide accurate and operationally useful precipitation predictions, overcoming the limitations of traditional approaches. These models show improved forecast quality, consistency, and value, making them a valuable tool for various sectors reliant on weather-dependent decision-making.
Article
Computer Science, Artificial Intelligence
Kevin Trebing et al.
Summary: Weather forecasting is mainly dominated by numerical weather prediction, which lacks the ability to make short-term forecasts using the latest information. Utilizing a data-driven neural network approach can produce accurate precipitation nowcasts. Experimental results show that the proposed SmaAt-UNet model performs comparably to other models in terms of prediction performance.
PATTERN RECOGNITION LETTERS
(2021)
Proceedings Paper
Computer Science, Artificial Intelligence
Haixu Wu et al.
Summary: This paper presents a novel video prediction framework, MotionRNN, which can simultaneously capture complex variations within motions and adapt to spacetime-varying scenarios. By designing MotionGRU unit and introducing Motion Highway, this framework significantly improves the ability to predict changeable motions.
2021 IEEE/CVF CONFERENCE ON COMPUTER VISION AND PATTERN RECOGNITION, CVPR 2021
(2021)
Article
Multidisciplinary Sciences
Yoo-Geun Ham et al.
Proceedings Paper
Computer Science, Artificial Intelligence
Du Tran et al.
2015 IEEE INTERNATIONAL CONFERENCE ON COMPUTER VISION (ICCV)
(2015)
Article
Computer Science, Artificial Intelligence
Z Wang et al.
IEEE TRANSACTIONS ON IMAGE PROCESSING
(2004)