Deep Convolutional Neural Networks (DCNN) have the ability to learn complex features and are thus widely used in the field of seismic signal denoising with low signal-to-noise ratio (SNR). However, the current convolutional deep network used for seismic signal noise reduction does not make full use of the feature information extracted from all convolution layers in the network, and thus cannot fit the seismic signal with high SNR. To deal with this problem, this paper proposes the DnRDB model, a convolutional deep network time-frequency domain seismic signal denoising model combined with residual dense blocks (RDB). The model is mainly composed of several RDB in series. The input of each convolution layer in each RDB module is formed by the output of all the previous convolution layers. Meanwhile, even if the number of layers is increased, the fusion of the seismic signal features learned by the RDB modules can still achieve full extraction of seismic signals. Furthermore, deepening the model structure by concatenating multiple RDB modules enables further useful feature information to be extracted, which improves the SNR of seismic signals. The DnRDB model was trained and tested using the Stanford Global Seismic Dataset. The experimental results show that the DnRDB model can effectively recover seismic signals and remove various forms of noise. Even in the case of high noise, the denoised signal still has a high SNR. When the DnRDB model is compared with other denoising approaches such as wavelet threshold, empirical mode decomposition, and different deep learning methods, the results indicate that it performs best overall in denoising the same segment of the noisy seismic signal; the denoised signal also has less waveform distortion. Use of the DnRDB model in subsequent seismic signal processing work indicates that it can help the phase recognition algorithm improve the accuracy of seismic recognition through noise reduction.

To address the problem of waveform distortion in the existing seismic signal denoising method when removing co-band noise, further improving the signal-to-noise ratio (SNR) of seismic signals and enhancing their quality, this paper designs a seismic co-band denoising model Atrous Residual Dense Block U-Net (ARDU), which uses a U-shaped convolutional neural network (U-Net) as a basic framework and combines atrous convolution and the residual dense block (RDB). In the ARDU model, atrous convolution is connected with residual dense blocks to form the feature extraction unit of the model encoder. Among them, the residual dense blocks can deepen the network’s depth and enhance the feature extraction ability of the network on the premise of mitigating the gradient-vanishing and gradient-exploding problem. Atrous convolution can enlarge receptive fields, reduce waveform distortion, and protect effective signals without increasing network parameters. To test the denoising performance of the ARDU model, the Stanford Global Seismic dataset was used to construct a training set and a test set and the model was trained and tested on it. The experimental results of the ARDU model for different types of seismic co-band noise showed that this model can effectively remove seismic co-band noise, protect effective signals, improve the SNR of seismic signals, and enhance the quality of seismic signals. To further verify the denoising effect of the model, this model was compared with the wavelet threshold denoising U-Net model and the denoising residual dense block (DnRDB) model, and the results showed that the ARDU model has the best SNR, r (correlation coefficient), and root-mean-square error (RMSE) and the least distortion of the seismic signal waveform.

Weather forecasting has been playing an important role in socio-economics. However, operational numerical weather prediction (NWP) is insufficiently accurate in terms of precipitation forecasting, especially for heavy rainfalls. Previous works on NWP bias correction utilizing deep learning (DL) methods mostly focused on a local region, and the China-wide precipitation forecast correction had not been attempted. Meanwhile, earlier studies imposed no particular focus on strong rainfalls despite their severe catastrophic impacts. In this study, we propose a DL model called weighted U-Net (WU-Net) that incorporates sample weights for various precipitation events to improve the forecasts of intensive precipitation in China. It is found that WU-Net can further improve the forecasting skill of heaviest rainfall comparing with the ordinary U-Net and ECMWF-IFS. Further analysis shows that this improvement increases with growing lead time, and distributes mainly in the eastern parts of China. This study suggests that a DL model considering the imbalance of the meteorological data could further improve the precipitation forecasting generated by numerical weather prediction.