Improving Nowcasting of Convective Development by Incorporating Polarimetric Radar Variables Into a Deep‐Learning Model

Pan, Xiang; Lu, Yinghui; Zhao, Kun; Huang, Hao; Wang, Mingjun; Chen, Haonan

doi:10.1029/2021gl095302

Cited by 70 publications

(47 citation statements)

References 52 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Zhou et al [18] used U-Net as a backbone model to leverage multisource data to forecast lightning in the next few hours. Pan et al [24] proposed FURENet, which incorporates late-fusion and channel-wise attention into U-Net, facilitating the usage of multiple polarimetric radar variables to benefit convective precipitation nowcasting. Ayzel et al [7] proposed RainNet based on U-Net for precipitation nowcasting, and thoroughly evaluated performance versus optical flow-based methods, with different precipitation thresholds and spectrum space analysis.…”

Section: Cnn-based Methodsmentioning

confidence: 99%

Towards a More Realistic and Detailed Deep-Learning-Based Radar Echo Extrapolation Method

Chen

Wang

et al. 2021

Remote Sensing

Self Cite

View full text Add to dashboard Cite

Deep-learning-based radar echo extrapolation methods have achieved remarkable progress in the precipitation nowcasting field. However, they suffer from a common notorious problem—they tend to produce blurry predictions. Although some efforts have been made in recent years, the blurring problem is still under-addressed. In this work, we propose three effective strategies to assist deep-learning-based radar echo extrapolation methods to achieve more realistic and detailed prediction. Specifically, we propose a spatial generative adversarial network (GAN) and a spectrum GAN to improve image fidelity. The spatial and spectrum GANs aim at penalizing the distribution discrepancy between generated and real images from the spatial domain and spectral domain, respectively. In addition, a masked style loss is devised to further enhance the details by transferring the detailed texture of ground truth radar sequences to extrapolated ones. We apply a foreground mask to prevent the background noise from transferring to the outputs. Moreover, we also design a new metric termed the power spectral density score (PSDS) to quantify the perceptual quality from a frequency perspective. The PSDS metric can be applied as a complement to other visual evaluation metrics (e.g., LPIPS) to achieve a comprehensive measurement of image sharpness. We test our approaches with both ConvLSTM baseline and U-Net baseline, and comprehensive ablation experiments on the SEVIR dataset show that the proposed approaches are able to produce much more realistic radar images than baselines. Most notably, our methods can be readily applied to any deep-learning-based spatiotemporal forecasting models to acquire more detailed results.

show abstract

Section: Cnn-based Methodsmentioning

confidence: 99%

Towards a More Realistic and Detailed Deep-Learning-Based Radar Echo Extrapolation Method

Chen

Wang

et al. 2021

Remote Sensing

Self Cite

View full text Add to dashboard Cite

show abstract

“…For example, the variational optical flow technique (VarFlow) [43] has been used to estimate the motion field in the latest precipitation nowcasting system which is operated by Hong Kong Observatory (HKO) [41]. To better utilize the vast amount of historical data, deep learning techniques such as 3-D CNN [44] and U-Net CNN [45], [46], have been applied. Formulating the precipitation nowcasting problem as an image-to-image translation problem and taking the spatiotemporal features into account, convolutional LSTM (Con-vLSTM) [47] and trajectory gated recurrent unit (TrajGRU) [48] have been proposed.…”

Section: A Related Workmentioning

confidence: 99%

A Spatio-Temporal Neural Network for Fine-Scale Wind Field Nowcasting Based on Lidar Observation

Gao

Shen

Zhou

et al. 2022

IEEE J. Sel. Top. Appl. Earth Observations Remote Sensing

View full text Add to dashboard Cite

Fine-scale wind field nowcasting is of great significance in air traffic management, power grid operation, and so on. In this article, an indirect wind field nowcasting scheme based on lidar observation is presented, which contains an encoderforecaster network based on the convolutional long short-term memory (ConvLSTM) with balanced structure and a mask branch. The proposed nowcasting network is trained and evaluated based on the lidar observations throughout 2020 at Hong Kong International Airport. Comprehensive comparison with nine methods including the widely used optical flow technique and classic neural network show the good performance of the new network. It can capture the spatio-temporal features in the lidar observations and obtain better nowcasting results up to 27 minutes with a resolution of 100 m. The nowcasting errors are smaller than the retrieval errors reported in recent literatures, demonstrating that the lidar observation nowcasting based on the new network can get fine-scale wind field nowcasting results with high efficiency.

show abstract

“…Inspired by the U-Net and SegNet series of deep learning models, Ayzel et al [30] proposed RainNet for radar precipitation near-prediction. Pan et al [31] proposed a novel deep learning model, FURENet, which is designed for extracting information from multiple input variables to make predictions. A GAN model proposed by Hu et al [32] has proven to be effective in overcoming the limitations of blurry predictions.…”

Section: Introductionmentioning

confidence: 99%

ConvLSTM Network-Based Rainfall Nowcasting Method with Combined Reflectance and Radar-Retrieved Wind Field as Inputs

et al. 2022

View full text Add to dashboard Cite

Strong convection nowcasting has been gaining importance in regional security, economic development, and water resource management. Rainfall nowcasting with strong timeliness needs to effectively forecast the intensity of rainfall in a local region in the short term. The forecast performance of traditional methods is limited. In this paper, a rainfall nowcasting model based on the Convolutional Long Short-Term Memory (ConvLSTM) is proposed. Combined reflectance (CR) and the retrieved wind field are selected as the key factors for prediction. The model considers the influence of water vapor content, transport, and change on rainfall by measuring CR and the retrieved wind field. Forecast results are compared to different models and different input data schemes. The analysis shows that the CSI scores of this method can reach about 0.8, which is 28% higher than that of the optical flow method. The ConvLSTM structure with spatial analysis and time memory can greatly enhance the predictive ability of the model, and the addition of wind field data also improves the forecasting performance of the model. Therefore, the idea of forecasting rainfall on the basis of water vapor content and water vapor transport is practical and effective, and the accuracy of a forecast can be improved by using ConvLSTM network to extract spatiotemporal features.

show abstract

Improving Nowcasting of Convective Development by Incorporating Polarimetric Radar Variables Into a Deep‐Learning Model

Cited by 70 publications

References 52 publications

Towards a More Realistic and Detailed Deep-Learning-Based Radar Echo Extrapolation Method

Towards a More Realistic and Detailed Deep-Learning-Based Radar Echo Extrapolation Method

A Spatio-Temporal Neural Network for Fine-Scale Wind Field Nowcasting Based on Lidar Observation

ConvLSTM Network-Based Rainfall Nowcasting Method with Combined Reflectance and Radar-Retrieved Wind Field as Inputs

Contact Info

Product

Resources

About