The seasonal signal and long-term trend in the height time series of 10 IGS sites in China are investigated in this paper. The offset detection and outlier removal as well as the removal of common mode error are perfo...The seasonal signal and long-term trend in the height time series of 10 IGS sites in China are investigated in this paper. The offset detection and outlier removal as well as the removal of common mode error are performed on the raw GPS time-series data developed by the Scripps Orbit and Permanent Array Center(SOPAC). The seasonal-trend decomposition procedure based on LOESS(STL) is utilized to extract precise seasonal signals, followed by an estimation of the long-term trend with the application of maximum likelihood estimation(MLE) to the seasonally adjusted time series. The Up-compo- nents of all sites are featured by obvious seasonal variations, with significant phase and amplitude modulation on some sites. After Kendall's tau test, a significant trend(99% confidence interval) for all sites is achieved. Furthermore, the trends at sites TCMS and TNML have significant changes at epochs 2009.5384 and 2009.1493(95% confidence interval), respectively, using the Breaks For Additive Seasonal and Trend test. Finally, the velocities and their uncertainties for all sites are estimated using MLE with the white noise plus flicker noise model. And the results are analyzed and compared with those announced by SOPAC. The results obtained in this paper have a higher precision than the SOPAC results.展开更多
Global navigation satellite system(GNSS)technique has irreplaceable advantages in the continuous monitoring of surface deformation.Reducing noise to improve the signal-to-noise ratio(SNR)and extract the concerned sign...Global navigation satellite system(GNSS)technique has irreplaceable advantages in the continuous monitoring of surface deformation.Reducing noise to improve the signal-to-noise ratio(SNR)and extract the concerned signals is of great significance.As an improved algorithm of empirical mode decomposition(EMD),complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)algorithm has better signal processing ability.Using the CEEMDAN algorithm,the height time series of 29GNSS stations in Chinese mainland were analyzed,and good denoising effects and extraction from periodic signals were achieved.The numerical results showed that the annual signal obtained with the CEEMDAN algorithm was significantly based on Lomb_Scargle spectrum analysis,and large differences in the long-term signals were found between the stations at different locations in Chinese mainland.With respect to data denoising,compared with the EMD and wavelet denoising algorithms,the CEEMDAN algorithm respectively improved the SNR by 29.35% and 36.54%,increased the correlation coefficient by 8.67% and 11.96%,and reduced root mean square error(RMSE)by 44.68% and 43.48%,indicating that the CEEMDAN algorithm had better denoising behavior than the other two algorithms.In addition,the results demonstrated that different denoising methods had little influence on estimating the annual vertical deformation velocity.The extraction of periodic signals showed that more components were retained by using the CEEMDAN algorithm than the EMD algorithm,which indicated that the CEEMDAN algorithm had advantages over frequency aliasing.In conclusion,the CEEMDAN algorithm was recommended for processing the GNSS height time series to analyze the vertical deformation due to its excellent features of denoising and the extraction of periodic signals.展开更多
The prediction of significant wave height(SWH)is crucial for managing wave energy.While many machine learning studies have focused on accurately predicting SWH values within hours in advance,the primary concern should...The prediction of significant wave height(SWH)is crucial for managing wave energy.While many machine learning studies have focused on accurately predicting SWH values within hours in advance,the primary concern should be given to the level of the wave height for real-world applications.In this paper,a classification framework for the time-series of SWH based on Transformer encoder(TF)and empirical mode decomposition(EMD)is developed,which can provide a lead time of 6 to 48 h with the fixed thresholds of 2 m for high level waves and 1.5 m for low level waves.The performance of this approach is compared to that of three mainstream algorithms with and without EMD features.Results from the datasets collected from buoy measurements in the Atlantic Ocean indicate that the optimal mean accuracy at a lead time of 6 h was 99.1%and the average training time was 75 s,demonstrating the accuracy and efficiency of this proposed model.This study provides valuable tools and references for real-world SWH prediction applications.展开更多
基金supported by the National High Technology Research and Development Program of China(Grant No.2013AA122501-1)the National Natural Science Foundation of China(Grant Nos.41374019,41020144004,41474015,41274045,41574010)Funded by State Key Laboratory of Geo-information Engineering(Grant No.SKLGIE2015-Z-1-1)
文摘The seasonal signal and long-term trend in the height time series of 10 IGS sites in China are investigated in this paper. The offset detection and outlier removal as well as the removal of common mode error are performed on the raw GPS time-series data developed by the Scripps Orbit and Permanent Array Center(SOPAC). The seasonal-trend decomposition procedure based on LOESS(STL) is utilized to extract precise seasonal signals, followed by an estimation of the long-term trend with the application of maximum likelihood estimation(MLE) to the seasonally adjusted time series. The Up-compo- nents of all sites are featured by obvious seasonal variations, with significant phase and amplitude modulation on some sites. After Kendall's tau test, a significant trend(99% confidence interval) for all sites is achieved. Furthermore, the trends at sites TCMS and TNML have significant changes at epochs 2009.5384 and 2009.1493(95% confidence interval), respectively, using the Breaks For Additive Seasonal and Trend test. Finally, the velocities and their uncertainties for all sites are estimated using MLE with the white noise plus flicker noise model. And the results are analyzed and compared with those announced by SOPAC. The results obtained in this paper have a higher precision than the SOPAC results.
基金supported by the National Natural Science Foundation of China(Grant No.42192535,42174012,42174101,41974023)the Open Fund of Hubei Luojia Laboratory(Grant No.S22H640201)。
文摘Global navigation satellite system(GNSS)technique has irreplaceable advantages in the continuous monitoring of surface deformation.Reducing noise to improve the signal-to-noise ratio(SNR)and extract the concerned signals is of great significance.As an improved algorithm of empirical mode decomposition(EMD),complete ensemble empirical mode decomposition with adaptive noise(CEEMDAN)algorithm has better signal processing ability.Using the CEEMDAN algorithm,the height time series of 29GNSS stations in Chinese mainland were analyzed,and good denoising effects and extraction from periodic signals were achieved.The numerical results showed that the annual signal obtained with the CEEMDAN algorithm was significantly based on Lomb_Scargle spectrum analysis,and large differences in the long-term signals were found between the stations at different locations in Chinese mainland.With respect to data denoising,compared with the EMD and wavelet denoising algorithms,the CEEMDAN algorithm respectively improved the SNR by 29.35% and 36.54%,increased the correlation coefficient by 8.67% and 11.96%,and reduced root mean square error(RMSE)by 44.68% and 43.48%,indicating that the CEEMDAN algorithm had better denoising behavior than the other two algorithms.In addition,the results demonstrated that different denoising methods had little influence on estimating the annual vertical deformation velocity.The extraction of periodic signals showed that more components were retained by using the CEEMDAN algorithm than the EMD algorithm,which indicated that the CEEMDAN algorithm had advantages over frequency aliasing.In conclusion,the CEEMDAN algorithm was recommended for processing the GNSS height time series to analyze the vertical deformation due to its excellent features of denoising and the extraction of periodic signals.
基金The financial support from the National Natural Science Foundation of China(No.61973208)is gratefully acknowledged.
文摘The prediction of significant wave height(SWH)is crucial for managing wave energy.While many machine learning studies have focused on accurately predicting SWH values within hours in advance,the primary concern should be given to the level of the wave height for real-world applications.In this paper,a classification framework for the time-series of SWH based on Transformer encoder(TF)and empirical mode decomposition(EMD)is developed,which can provide a lead time of 6 to 48 h with the fixed thresholds of 2 m for high level waves and 1.5 m for low level waves.The performance of this approach is compared to that of three mainstream algorithms with and without EMD features.Results from the datasets collected from buoy measurements in the Atlantic Ocean indicate that the optimal mean accuracy at a lead time of 6 h was 99.1%and the average training time was 75 s,demonstrating the accuracy and efficiency of this proposed model.This study provides valuable tools and references for real-world SWH prediction applications.
